From psatpute at redhat.com Thu Jan 2 10:05:06 2014 From: psatpute at redhat.com (Pravin Satpute) Date: Thu, 02 Jan 2014 15:35:06 +0530 Subject: [Lohit-devel-list] [Announce] Released 2.92.2 version of the Lohit Gujarati (Lohit2 project) Message-ID: <52C539D2.6010205@redhat.com> Hi All, First of all happy new year !! There was one interesting bug [1] got reported against 2.92.1 release of Lohit Gujarati. After lots of debug found it was introduced due to our recent additions of extended i-matras and corresponding rules. We have resolved this issue. Due to high severity of bug we are doing this minor release. I believe Lohit Gujarati is stable now and we can consider it in low maintenance mode now. Updated lohit project page [2] for download details. Source tarball link [3], TTF tarball link [4] and webfonts format for Lohit is at [5]. I will soon build this for Fedora. Feedback and suggestions for improvements are welcome. Regards, Pravin Satpute 1. https://github.com/pravins/lohit/issues/31 2. https://fedorahosted.org/lohit 3. https://fedorahosted.org/releases/l/o/lohit/lohit-gujarati-2.92.2.tar.gz 4. https://fedorahosted.org/releases/l/o/lohit/lohit-gujarati-ttf-2.92.2.tar.gz 5. https://fedorahosted.org/releases/l/o/lohit/lohit-gujarati-web-2.92.2.tar.gz 6. https://github.com/pravins/lohit/issues/new From skore at redhat.com Fri Jan 3 09:14:14 2014 From: skore at redhat.com (Sneha Kore) Date: Fri, 3 Jan 2014 04:14:14 -0500 (EST) Subject: [Lohit-devel-list] Query for Malayalam Font In-Reply-To: <1550311442.35644975.1388734095600.JavaMail.root@redhat.com> Message-ID: <1919692090.35659668.1388740454591.JavaMail.root@redhat.com> Hi all, As you all know, we had started to work for Lohit Malayalam. Presently we are at the final stage of testing. For Malayalam development we followed TDIL http://tdil.mit.gov.in/pdf/Vishwabharat/tdil-april-2002.zip. As per the Malayalam script details given by TDIL , It says, RAKAR is formed when Malayalam letter RA(0D30) or RRA(0D31) is at the end of consonant/consonant cluster, a special sign is proposed to the consonant. but the query arising is, weather RAKAR should be formed for lettr RRA (0D31) ?? If yes, then we supposed to file a bug on harfbuzz since this is rendering issue. The reference regarding the same can be useful. Thanks & Regards, Sneha Kore -------------- next part -------------- A non-text attachment was scrubbed... Name: TDIL-Malayalam.png Type: image/png Size: 358804 bytes Desc: not available URL: From pravin.d.s at gmail.com Fri Jan 3 09:28:03 2014 From: pravin.d.s at gmail.com (pravin.d.s at gmail.com) Date: Fri, 3 Jan 2014 14:58:03 +0530 Subject: [Lohit-devel-list] Query for Malayalam Font In-Reply-To: <1919692090.35659668.1388740454591.JavaMail.root@redhat.com> References: <1550311442.35644975.1388734095600.JavaMail.root@redhat.com> <1919692090.35659668.1388740454591.JavaMail.root@redhat.com> Message-ID: On 3 January 2014 14:44, Sneha Kore wrote: > Hi all, > > As you all know, we had started to work for Lohit Malayalam. Presently we > are at the final stage of testing. For Malayalam development we followed > TDIL http://tdil.mit.gov.in/pdf/Vishwabharat/tdil-april-2002.zip. As per > the Malayalam script details given by TDIL , It says, > RAKAR is formed when Malayalam letter RA(0D30) or RRA(0D31) is at the end > of consonant/consonant cluster, a special sign is proposed to the > consonant. but the query arising is, weather RAKAR should be formed for > lettr RRA (0D31) ?? If yes, then we supposed to file a bug on harfbuzz > since this is rendering issue. The reference regarding the same can be > useful. > I am doubtful that people might be using this in iscii and same is proposed by TDIL. forming same ligature by both U+0D30 and U+0D31 looks doubtful to me as well. Unicode says nothing on this. Regards, Pravin Satpute -------------- next part -------------- An HTML attachment was scrubbed... URL: From samjnaa at gmail.com Sat Jan 4 13:38:54 2014 From: samjnaa at gmail.com (Shriramana Sharma) Date: Sat, 4 Jan 2014 19:08:54 +0530 Subject: [Lohit-devel-list] Query for Malayalam Font In-Reply-To: References: <1550311442.35644975.1388734095600.JavaMail.root@redhat.com> <1919692090.35659668.1388740454591.JavaMail.root@redhat.com> Message-ID: On Fri, Jan 3, 2014 at 2:58 PM, pravin.d.s at gmail.com wrote: > > I am doubtful that people might be using this in iscii and same is proposed > by TDIL. > forming same ligature by both U+0D30 and U+0D31 looks doubtful to me as > well. Unicode says nothing on this. Unicode says in general (TUS 6.2 p 286, p 316 of PDF) that C + VIRAMA + RA forms C2-conjoining form of RA. It is inadvisable to have both KA + VIRAMA + RA/RRA to render as the same as it would result in confusability issues in IDNs and hence phishing attacks. -- Shriramana Sharma ???????????? ???????????? From pravin.d.s at gmail.com Mon Jan 6 05:37:35 2014 From: pravin.d.s at gmail.com (pravin.d.s at gmail.com) Date: Mon, 6 Jan 2014 11:07:35 +0530 Subject: [Lohit-devel-list] Query for Malayalam Font In-Reply-To: References: <1550311442.35644975.1388734095600.JavaMail.root@redhat.com> <1919692090.35659668.1388740454591.JavaMail.root@redhat.com> Message-ID: On 4 January 2014 19:08, Shriramana Sharma wrote: > On Fri, Jan 3, 2014 at 2:58 PM, pravin.d.s at gmail.com > wrote: > > > > I am doubtful that people might be using this in iscii and same is > proposed > > by TDIL. > > forming same ligature by both U+0D30 and U+0D31 looks doubtful to me as > > well. Unicode says nothing on this. > > Unicode says in general (TUS 6.2 p 286, p 316 of PDF) that C + VIRAMA > + RA forms C2-conjoining form of RA. It is inadvisable to have both KA > + VIRAMA + RA/RRA to render as the same as it would result in > confusability issues in IDNs and hence phishing attacks. > Indeed. We should not repeat mistake done earlier with eyelash-reph [1] of Devanagari. I will reply to TDIL for fixing it in there next release. Thanks, Pravin Satpute 1. Eyelash reph can be form with both U+0930 and U+0931. -------------- next part -------------- An HTML attachment was scrubbed... URL: From bugzilla at redhat.com Thu Jan 9 09:35:24 2014 From: bugzilla at redhat.com (bugzilla at redhat.com) Date: Thu, 09 Jan 2014 09:35:24 +0000 Subject: [Lohit-devel-list] [Bug 1016989] [ml_IN] non-standard sequence sequence for stacked chillu-N and RRA in Lohit Malayalam font should be removed In-Reply-To: References: Message-ID: https://bugzilla.redhat.com/show_bug.cgi?id=1016989 Pravin Satpute changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |lohit-devel-list at redhat.com --- Comment #1 from Pravin Satpute --- This bug report is very clear and i agree with you. but surprisingly some fonts around does not follow it. Dunno why people are considering U+0D28 and U+0D7B as equivalent. In reality u0D28 + u0D4D + u200D = u0D7B I hope people around are not using this u0D28 instead of u0D7b and already created some data. -- You are receiving this mail because: You are on the CC list for the bug. Unsubscribe from this bug https://bugzilla.redhat.com/token.cgi?t=rsie9MTVz3&a=cc_unsubscribe From pravin.d.s at gmail.com Fri Jan 10 10:15:00 2014 From: pravin.d.s at gmail.com (pravin.d.s at gmail.com) Date: Fri, 10 Jan 2014 15:45:00 +0530 Subject: [Lohit-devel-list] Handling Malayalam "NTA" issue for Lohit2 Message-ID: Hi All, We are working on lohit2[1] project, whose plan is to create standard and reusable open type tables with additional improvement. Lohit as a default system fonts in most of the open source distros always follow standard around language technology. (Font specification, Storage, Guideline related to Languages) Recently we started working on Lohit Malayalam font [2] with some planned improvement and came across couple of bugs related [3][4] with well know "NTA" issue introduced during the addition of Atomic chillu characters in Unicode 5.1 Now dilemma is number of users already using * A. u0D28 + u0D4D + u0D31 for getting NTA character even before Unicode 5.1 * * B. But Unicode from 5.1 onward says (TUS 6.2 chapter 9.9 p 321) use u0D7B + u0D4D + u0D31 for getting same "NTA" * In my humble opinion here one thing is very clear that Unicode forgot to add normalization (backward compatibility) for newly added sequence in (B). Still i have not seen any improvement in it from long time. Now dilemma with lohit2 development is - Lohit 1 is supporting sequence (A) from long time (even before Unicode 5.1), so for the backward compatibility lohit2 should support the same. - Since Lohit follows standards, it is important to support sequence (B) for following Unicode 6.3. But following Unicode 6.3 in this case clearly invites dual encoding without any normalization rules handy. Good documentation on NTA issues is available at [5] Presently i am in favour of not supporting Unicode defined sequence (B) in lohit2 and keep on using (A) which is used in Lohit fonts family from long time. Please let me know your view on it. Is there any chance of getting this mention in Unicode chapter 9? is there any chance of Normalization rule for this? Regards, Pravin Satpute 1. http://pravin-s.blogspot.in/2013/08/project-creating-standard-and-reusable.html 2. http://pravin-s.blogspot.in/2013/12/lohit2-lohit-malayalam-development-plans.html 3. https://bugzilla.redhat.com/show_bug.cgi?id=1016984 4. https://bugzilla.redhat.com/show_bug.cgi?id=1016989 5. http://thottingal.in/documents/Malayalam-NTA.pdf -------------- next part -------------- An HTML attachment was scrubbed... URL: From samjnaa at gmail.com Fri Jan 10 12:24:46 2014 From: samjnaa at gmail.com (Shriramana Sharma) Date: Fri, 10 Jan 2014 17:54:46 +0530 Subject: [Lohit-devel-list] Handling Malayalam "NTA" issue for Lohit2 In-Reply-To: References: Message-ID: On Fri, Jan 10, 2014 at 3:45 PM, pravin.d.s at gmail.com wrote: > In my humble opinion here one thing is very clear that Unicode forgot to > add normalization (backward compatibility) for newly added sequence in (B). Dear Pravin, If by normalization you mean http://www.unicode.org/glossary/#normalization -- then it is not possible in this case since the individually encoded chillus do not have canonical decomposition to their related consonants. Indeed, that would defeat the purpose of the separate encoding, which was to provide semantically distinct chillus! The recent additional chillus trickling into the standard seems to indicate that one should have encoded a CHILLU MARKER back then, but there's no going back now, so chillus galore! ;-) On a more serious note, I think it is important to adhere to the standard, as it is good for you in the long run even though it is difficult at first. If you delay the adoption of the standard, it only gets all the harder as time passes, since in the interim even more people continue to assume the old behaviour... -- Shriramana Sharma ???????????? ???????????? From paivakil at gmail.com Fri Jan 10 17:46:30 2014 From: paivakil at gmail.com (Mahesh T. Pai) Date: Fri, 10 Jan 2014 23:16:30 +0530 Subject: [Lohit-devel-list] Handling Malayalam "NTA" issue for Lohit2 In-Reply-To: References: Message-ID: <20140110174630.GA18104@localhost> pravin.d.s at gmail.com said on Fri, Jan 10, 2014 at 03:45:00PM +0530,: - Lohit 1 is supporting sequence (A) from long time (even before > Unicode 5.1), so for the backward compatibility lohit2 should support the > same. > I believe thet the UTC wanted to maintain compatibility with some _beta_ version of Microsoft's some software in making the choice that it did regarding the /nta/ sequence. > Presently i am in favour of not supporting Unicode defined > sequence (B) in lohit2 and keep on using (A) which is used in Lohit > fonts family from long time. Allow me to go on a nostalgia trip. Almost a decade back, the then SMC team came accross what was obvious lack of clarity in the UTS. They decided, against my advise, to follow the suggestions in OpenType definition. To be fair, then, I had no alternative to offer, except not to implement the suggestion in the OpenType pages. Microsoft ultimately waited for some clarity in the UTS before implementing anything. and the communimity efforts went (mostly) in vain. Right now, given a choice between supporting legacy data and standards, I will choose the latter, with some kind of jugaad based on the PUA / glyph name to enable support for legacy data. Not the ideal situation, but when politics get the uppoer hand over merits, efficiency and appropriateness always takes a backseat. -- Mahesh T. Pai || free - (adj) able to act at will; not hampered; not under compulsion or restraint; free from obligations or duties; not bound to servitude; at liberty. From pravin.d.s at gmail.com Mon Jan 13 06:04:33 2014 From: pravin.d.s at gmail.com (pravin.d.s at gmail.com) Date: Mon, 13 Jan 2014 11:34:33 +0530 Subject: [Lohit-devel-list] Handling Malayalam "NTA" issue for Lohit2 In-Reply-To: References: Message-ID: On 10 January 2014 17:54, Shriramana Sharma wrote: > On Fri, Jan 10, 2014 at 3:45 PM, pravin.d.s at gmail.com > wrote: > > In my humble opinion here one thing is very clear that Unicode > forgot to > > add normalization (backward compatibility) for newly added sequence in > (B). > > Dear Pravin, > > If by normalization you mean > http://www.unicode.org/glossary/#normalization -- then it is not > possible in this case since the individually encoded chillus do not > have canonical decomposition to their related consonants. Indeed, that > would defeat the purpose of the separate encoding, which was to > provide semantically distinct chillus! > Ok not normalization but at least Unicode should mention old habit of writing NTA and new with addition of atomic chillu. It will definitely help people working on NLP to handle data having these two different sequence. > > On a more serious note, I think it is important to adhere to the > standard, as it is good for you in the long run even though it is > difficult at first. If you delay the adoption of the standard, it only > gets all the harder as time passes, since in the interim even more > people continue to assume the old behaviour... > >From font perspective if we consider there is NTA sequence is available in both form (A) & (B) in data around. We have to add required rules for both way. Unfortunately in this case Unicode has not consider for backward compatibility but at least Lohit project definitely consider it. So to be in safer side now i am fever of having both rules in font. Regards, Pravin Satpute -------------- next part -------------- An HTML attachment was scrubbed... URL: From pravin.d.s at gmail.com Mon Jan 13 06:28:52 2014 From: pravin.d.s at gmail.com (pravin.d.s at gmail.com) Date: Mon, 13 Jan 2014 11:58:52 +0530 Subject: [Lohit-devel-list] Handling Malayalam "NTA" issue for Lohit2 In-Reply-To: <20140110174630.GA18104@localhost> References: <20140110174630.GA18104@localhost> Message-ID: On 10 January 2014 23:16, Mahesh T. Pai wrote: > pravin.d.s at gmail.com said on Fri, Jan 10, 2014 at 03:45:00PM +0530,: > - Lohit 1 is supporting sequence (A) from long time (even before > > Unicode 5.1), so for the backward compatibility lohit2 should support > the > > same. > > > > I believe thet the UTC wanted to maintain compatibility with some > _beta_ version of Microsoft's some software in making the choice that > it did regarding the /nta/ sequence. > > > > Presently i am in favour of not supporting Unicode defined > > sequence (B) in lohit2 and keep on using (A) which is used in Lohit > > fonts family from long time. > > Allow me to go on a nostalgia trip. Almost a decade back, the then SMC > team came accross what was obvious lack of clarity in the UTS. They > decided, against my advise, to follow the suggestions in OpenType > definition. To be fair, then, I had no alternative to offer, except > not to implement the suggestion in the OpenType pages. Microsoft > ultimately waited for some clarity in the UTS before implementing > anything. and the communimity efforts went (mostly) in vain. > I was wondering how ISCII was handling this. > > Right now, given a choice between supporting legacy data and > standards, I will choose the latter, with some kind of jugaad based on > the PUA / glyph name to enable support for legacy data. > Yeah, as said above will support both legacy and standard sequence. > > Not the ideal situation, but when politics get the uppoer hand over > merits, efficiency and appropriateness always takes a backseat. > That is pain point of standardization activities. Thanks & Regards, Pravin Satpute -------------- next part -------------- An HTML attachment was scrubbed... URL: From bugzilla at redhat.com Mon Jan 13 09:08:26 2014 From: bugzilla at redhat.com (bugzilla at redhat.com) Date: Mon, 13 Jan 2014 09:08:26 +0000 Subject: [Lohit-devel-list] [Bug 1016989] [ml_IN] non-standard sequence sequence for stacked chillu-N and RRA in Lohit Malayalam font should be removed In-Reply-To: References: Message-ID: https://bugzilla.redhat.com/show_bug.cgi?id=1016989 Pravin Satpute changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |CLOSED CC| |snekore at gmail.com Resolution|--- |NOTABUG Last Closed| |2014-01-13 04:08:26 --- Comment #2 from Pravin Satpute --- 1. There is already good amount of data available using this legacy sequence. 2. Same time, this sequence is supported in Lohit from long time. Removing it will introduce regression. Lohit2 will also support this sequence with standard one given in Unicode. Closing this as a Not a Bug. Feel free to reopen if more issues. -- You are receiving this mail because: You are on the CC list for the bug. Unsubscribe from this bug https://bugzilla.redhat.com/token.cgi?t=IjRb8ZARBB&a=cc_unsubscribe From cibucj at gmail.com Mon Jan 13 06:32:16 2014 From: cibucj at gmail.com (=?UTF-8?B?4LS44LS/4LSs4LWBIOC0uOC0vyDgtJzgtYY=?=) Date: Sun, 12 Jan 2014 22:32:16 -0800 Subject: [Lohit-devel-list] Handling Malayalam "NTA" issue for Lohit2 In-Reply-To: References: Message-ID: In fact, there is one more sequence to consider. Kartika in Windows follows for NTA. However, the existing data in that sequence is quite less. In case, Chillus standard is asking display software to be prepared for data in both sequences. I agree, it could document NTA's legacy Vs standard sequences, likewise. 2014/1/12 pravin.d.s at gmail.com > > > > On 10 January 2014 17:54, Shriramana Sharma wrote: > >> On Fri, Jan 10, 2014 at 3:45 PM, pravin.d.s at gmail.com >> wrote: >> > In my humble opinion here one thing is very clear that Unicode >> forgot to >> > add normalization (backward compatibility) for newly added sequence in >> (B). >> >> Dear Pravin, >> >> If by normalization you mean >> http://www.unicode.org/glossary/#normalization -- then it is not >> possible in this case since the individually encoded chillus do not >> have canonical decomposition to their related consonants. Indeed, that >> would defeat the purpose of the separate encoding, which was to >> provide semantically distinct chillus! >> > > Ok not normalization but at least Unicode should mention old habit of > writing NTA and new with addition of atomic chillu. It will definitely help > people working on NLP to handle data having these two different sequence. > > >> >> On a more serious note, I think it is important to adhere to the >> standard, as it is good for you in the long run even though it is >> difficult at first. If you delay the adoption of the standard, it only >> gets all the harder as time passes, since in the interim even more >> people continue to assume the old behaviour... >> > > From font perspective if we consider there is NTA sequence is available in > both form (A) & (B) in data around. We have to add required rules for both > way. Unfortunately in this case Unicode has not consider for backward > compatibility but at least Lohit project definitely consider it. > > So to be in safer side now i am fever of having both rules in font. > > Regards, > Pravin Satpute > > > > _______________________________________________ > Indic mailing list > Indic at unicode.org > http://unicode.org/mailman/listinfo/indic > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From psatpute at redhat.com Wed Jan 22 13:10:32 2014 From: psatpute at redhat.com (Pravin Satpute) Date: Wed, 22 Jan 2014 18:40:32 +0530 Subject: [Lohit-devel-list] Lohit2: Automated testing support to avoid regressions in Lohit Message-ID: <52DFC348.90502@redhat.com> Hi All, Life was easy from testing perspective till we were working on Devanagari and Gujarati scripts due to bit familiar shapes :) It became bit tricky from the time we started working on Malayalam script ;) Still thanks to Unicode and Reference document from TDIL and continuous help from Ani. We started facing need for automated testing. I remember we discussed during the Open source language summit 2013 regarding automated testing and Santhosh pointed me to GSoc project for automated rendering test cases. I liked the idea of this project i.e. keeping expected hb-shape output in standard file and then after each modification comparing output with standard file. We are doing same in Lohit with small python code done. [1] Regressions are not new for Indian scripts :) I am sure with automated test module will help Lohit fonts in future from introducing unexpected regressions. Cheers, Pravin Satpute 1. https://raw.github.com/pravins/lohit/master/scripts/auto_test.py From kamathvasudev at gmail.com Wed Jan 22 16:32:29 2014 From: kamathvasudev at gmail.com (Vasudev Kamath) Date: Wed, 22 Jan 2014 22:02:29 +0530 Subject: [Lohit-devel-list] Lohit2: Automated testing support to avoid regressions in Lohit In-Reply-To: <52DFC348.90502@redhat.com> (Pravin Satpute's message of "Wed, 22 Jan 2014 18:40:32 +0530") References: <52DFC348.90502@redhat.com> Message-ID: <87r480dlk2.fsf@vasudev.homelinux.net> Pravin Satpute writes: Hi Pravin, > Hi All, > > Life was easy from testing perspective till we were working on > Devanagari and Gujarati scripts due to bit familiar shapes :) > It became bit tricky from the time we started working on Malayalam > script ;) Still thanks to Unicode and Reference document from TDIL and > continuous help from Ani. > > We started facing need for automated testing. I remember we > discussed during the Open source language summit 2013 regarding > automated testing and Santhosh pointed me to GSoc project for automated > rendering test cases. > > I liked the idea of this project i.e. keeping expected hb-shape > output in standard file and then after each modification comparing > output with standard file. > > We are doing same in Lohit with small python code done. [1] I'm just wondering why you are re-inventing the wheel. Why can't you use automated-redner-testing tool which is already on github? [*] [*] https://github.com/nandajavarma/Automated-Rendering-Testing/ You just need to add needed data files for each language and add it to some CI system like travis. -- Vasudev Kamath http://copyninja.info Connect on ~friendica: copyninja@{frndk.de | vasudev.homelinux.net} IRC nick: copyninja | vasudev {irc.oftc.net | irc.freenode.net} GPG Key: C517 C25D E408 759D 98A4 C96B 6C8F 74AE 8770 0B7E -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 818 bytes Desc: not available URL: From pravin.d.s at gmail.com Wed Jan 22 18:13:02 2014 From: pravin.d.s at gmail.com (pravin.d.s at gmail.com) Date: Wed, 22 Jan 2014 23:43:02 +0530 Subject: [Lohit-devel-list] Lohit2: Automated testing support to avoid regressions in Lohit In-Reply-To: <87r480dlk2.fsf@vasudev.homelinux.net> References: <52DFC348.90502@redhat.com> <87r480dlk2.fsf@vasudev.homelinux.net> Message-ID: On 22 January 2014 22:02, Vasudev Kamath wrote: > Pravin Satpute writes: > > Hi Pravin, > > > Hi All, > > > > Life was easy from testing perspective till we were working on > > Devanagari and Gujarati scripts due to bit familiar shapes :) > > It became bit tricky from the time we started working on Malayalam > > script ;) Still thanks to Unicode and Reference document from TDIL and > > continuous help from Ani. > > > > We started facing need for automated testing. I remember we > > discussed during the Open source language summit 2013 regarding > > automated testing and Santhosh pointed me to GSoc project for automated > > rendering test cases. > > > > I liked the idea of this project i.e. keeping expected hb-shape > > output in standard file and then after each modification comparing > > output with standard file. > > > > We are doing same in Lohit with small python code done. [1] > > I'm just wondering why you are re-inventing the wheel. Why can't you use > automated-redner-testing tool which is already on github? [*] > > [*] https://github.com/nandajavarma/Automated-Rendering-Testing/ > > You just need to add needed data files for each language and add it to > some CI system like travis. > Good question. Present requirements of lohit are very minimal, it is just string comparison of command prompt output with existing file using python. I will definitely consider using Automated-Rendering-Testingin future for some more complex tasks and features getting implemented there. Regards, Pravin Satpute -------------- next part -------------- An HTML attachment was scrubbed... URL: From dave at lab6.com Thu Jan 23 01:02:43 2014 From: dave at lab6.com (Dave Crossland) Date: Thu, 23 Jan 2014 01:02:43 +0000 Subject: [Lohit-devel-list] Lohit2: Automated testing support to avoid regressions in Lohit In-Reply-To: References: <52DFC348.90502@redhat.com> <87r480dlk2.fsf@vasudev.homelinux.net> Message-ID: On 22 January 2014 18:13, pravin.d.s at gmail.com wrote: >> [*] https://github.com/nandajavarma/Automated-Rendering-Testing/ >> >> You just need to add needed data files for each language and add it to >> some CI system like travis. > > > Good question. > Present requirements of lohit are very minimal, it is just string comparison > of command prompt output with existing file using python. http://github.com/xen/fontbakery is the font CI system I have been working on with a Ukrainian contractor for about a year. There are a few screencasts explaining it at http://fontbakery.tumblr.com/ and I hope to update with another next week. Very happy to learn of nandajavarma's project!! :) -- Cheers Dave From psatpute at redhat.com Tue Jan 28 09:37:51 2014 From: psatpute at redhat.com (Pravin Satpute) Date: Tue, 28 Jan 2014 15:07:51 +0530 Subject: [Lohit-devel-list] Improvements going to happen in Lohit Gurmukhi under lohit2 project Message-ID: <52E77A6F.3030305@redhat.com> Hi All, On the way to do Alpha release of Lohit Malayalam. Lets look towards Lohit Punjabi from improvement side in Lohit2 project[1] One of the main change we are going to do in Lohit2 is we are going to call it "Lohit Gurmukhi" rather than Punjabi. Fonts are for writing script so it is good to call them by script name rather than language. Same time it is also less confusing for languages using multiple scripts. Punjabi language uses Arabic as well Gurmukhi. Bug[2] is already open for this change from A.S Alam. Other regular improvements of Lohit2 includes * Following AGL specification (We are just following syntax and not naming stuff yet. It will happen with script improvement for generating sfd to ttf) * Feature file separate for flexibility and reusability. * Complete cleanup of existing Open type tables * Supporting both "guru" and "gur2" open type specifications. * Thorough testing with Harfbuzz and Uniscribe (WinXp, W7 and W8) * Fixing existing open bugs of Lohit Punjabi [3] Feel free to suggest any other possible improvement in Lohit Gurmukhi. Expecting to do Alpha release in mid Feb. The development has already been started at Lohit2 https://github.com/pravins/lohit2/tree/master/punjabi Best Regards, Pravin Satpute 1. http://pravin-s.blogspot.in/2013/08/project-creating-standard-and-reusable.html 2. https://bugzilla.redhat.com/show_bug.cgi?id=1047373 3. https://bugzilla.redhat.com/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&classification=Fedora&component=lohit-punjabi-fonts&list_id=2147810&product=Fedora&query_format=advanced -------------- next part -------------- An HTML attachment was scrubbed... URL: From pravin.d.s at gmail.com Wed Jan 29 05:10:11 2014 From: pravin.d.s at gmail.com (pravin.d.s at gmail.com) Date: Wed, 29 Jan 2014 10:40:11 +0530 Subject: [Lohit-devel-list] Help required for Lohit Gujarati issue on Android 4.2.2 Message-ID: Hi All, Facing an interesting issue with Lohit Gujarati 2.92.2 on Android 4.2.2. Issue is matras do not get properly attached to the consonant ka (?) and cha (?). They work fine at all other consonants. Issue is reported on github [1] I did thorough analysis of Lohit Gujarati but still not getting any clue about why this is happening. On the other hand i feel this is issue from rendering engine side as Lohit Gujarati works fine on Desktop side. Reporter says its working fine with Samyak fonts and Shruti fonts. Any quick guessing? Best Regards, Pravin Satpute 1. https://github.com/pravins/lohit/issues/34 -------------- next part -------------- An HTML attachment was scrubbed... URL: From psatpute at redhat.com Thu Jan 30 13:21:03 2014 From: psatpute at redhat.com (Pravin Satpute) Date: Thu, 30 Jan 2014 18:51:03 +0530 Subject: [Lohit-devel-list] Announcing alpha release of lohit-malayalam 2.91.0 from lohit2 project Message-ID: <52EA51BF.40301@redhat.com> Hi All, Glad to announce alpha release of lohit-malayalam 2.91.0 from lohit2 project[1]. As we planned earlier [2] indeed this is lightest Malayalam modern script font around with just 214 glyphs and 11 open type table. We have done following improvements in Alpha version over the 2.5.4 release. * Reduced number of glyphs nearly by 25% * Open type Looksup are reduced by 45% * Supporting mlm2 and mlym specifications. * Testing on Harfbuzz NG and Uniscribe (W8) * Followed AGL syntax with User friendly glyph names. * Improved GASP table (grid-fitting) * TTF build from feature file. Updated lohit project page [3] for download details. Source tarball link [4], TTF tarball link [5] and webfonts format for Lohit is at [6]. I would like to say thanks to Ani for her help in testing stuff, Shriramana for reporting important bugs from standard perspective and Sneha for making this happen :) As this is alpha release need your help for testing it and making it more perfect. I am still doubtful that we might have missed some important exceptions of Malayalam script so looking forward to fix them in coming releases. By next week this release will be available in Fedora 20 and Rawhide. Cheers, Pravin Satpute 1. http://pravin-s.blogspot.in/2013/08/project-creating-standard-and-reusable.html 2. https://www.redhat.com/archives/lohit-devel-list/2013-December/msg00005.html 3. https://fedorahosted.org/lohit 4. https://fedorahosted.org/releases/l/o/lohit/lohit-malayalam-2.91.0.tar.gz 5. https://fedorahosted.org/releases/l/o/lohit/lohit-malayalam-ttf-2.91.0.tar.gz 6. https://fedorahosted.org/releases/l/o/lohit/lohit-malayalam-web-2.91.0.tar.gz 7. https://github.com/pravins/lohit/issues/new -------------- next part -------------- An HTML attachment was scrubbed... URL: From samjnaa at gmail.com Thu Jan 30 14:17:08 2014 From: samjnaa at gmail.com (Shriramana Sharma) Date: Thu, 30 Jan 2014 19:47:08 +0530 Subject: [Lohit-devel-list] Please don't replace stable font links by unstable ones Message-ID: Dear Pravin, When you are posting development releases on the http://fedorahosted.org/lohit/ page, please clearly differentiate the devel versions and do not include them or replace them in the stable releases list. As you know, all projects always separate the two so users who are not ready to use alpha/beta releases can use the latest known stable version. Right now stable versions of Devanagari (including for Marathi, Nepali), Gujarati and Malayalam are not accessible from the fedorahosted page. Please fix this. Another idea would be to make a subgroup of fonts that are really of the same script, Assamese/Bengali, Devanagari/Marathi/Nepali and Tamil/Tamil Classical, but that can be done later. First please restore the links to stable releases of fonts mentioned above. -- Shriramana Sharma ???????????? ???????????? From pravin.d.s at gmail.com Thu Jan 30 15:28:11 2014 From: pravin.d.s at gmail.com (pravin.d.s at gmail.com) Date: Thu, 30 Jan 2014 20:58:11 +0530 Subject: [Lohit-devel-list] Please don't replace stable font links by unstable ones In-Reply-To: References: Message-ID: On 30 January 2014 19:47, Shriramana Sharma wrote: > Dear Pravin, > > When you are posting development releases on the > http://fedorahosted.org/lohit/ page, please clearly differentiate the > devel versions and do not include them or replace them in the stable > releases list. As you know, all projects always separate the two so > users who are not ready to use alpha/beta releases can use the latest > known stable version. > > Right now stable versions of Devanagari (including for Marathi, > Nepali), Gujarati and Malayalam are not accessible from the > fedorahosted page. Please fix this. > Agree with you. Fixed this. > > Another idea would be to make a subgroup of fonts that are really of > the same script, Assamese/Bengali, Devanagari/Marathi/Nepali and > Tamil/Tamil Classical, but that can be done later. First please > restore the links to stable releases of fonts mentioned above. > Yes, this is also good idea it will give good picture of number of script supported by Lohit and languages covered in each script. Will do this later on. Regards, Pravin Satpute -------------- next part -------------- An HTML attachment was scrubbed... URL: From samjnaa at gmail.com Thu Jan 30 16:31:11 2014 From: samjnaa at gmail.com (Shriramana Sharma) Date: Thu, 30 Jan 2014 22:01:11 +0530 Subject: [Lohit-devel-list] Please don't replace stable font links by unstable ones In-Reply-To: References: Message-ID: On Thu, Jan 30, 2014 at 8:58 PM, pravin.d.s at gmail.com wrote: >> >> Right now stable versions of Devanagari (including for Marathi, >> Nepali), Gujarati and Malayalam are not accessible from the >> fedorahosted page. Please fix this. > > Agree with you. Fixed this. Checked just now -- only Lohit Malayalam alpha is given separately. Please list all alphas including Devanagari (+variants) and Gujarati separately from stables. Thanks. -- Shriramana Sharma ???????????? ???????????? From pravin.d.s at gmail.com Fri Jan 31 05:18:14 2014 From: pravin.d.s at gmail.com (pravin.d.s at gmail.com) Date: Fri, 31 Jan 2014 10:48:14 +0530 Subject: [Lohit-devel-list] Please don't replace stable font links by unstable ones In-Reply-To: References: Message-ID: On 30 January 2014 22:01, Shriramana Sharma wrote: > On Thu, Jan 30, 2014 at 8:58 PM, pravin.d.s at gmail.com > wrote: > >> > >> Right now stable versions of Devanagari (including for Marathi, > >> Nepali), Gujarati and Malayalam are not accessible from the > >> fedorahosted page. Please fix this. > > > > Agree with you. Fixed this. > > Checked just now -- only Lohit Malayalam alpha is given separately. > Please list all alphas including Devanagari (+variants) and Gujarati > separately from stables. Thanks. > Lohit Devanagari (+variants) and Gujarati are now in stable stage. Regards, Pravin Satpute -------------- next part -------------- An HTML attachment was scrubbed... URL: From pravin.d.s at gmail.com Fri Jan 31 08:14:08 2014 From: pravin.d.s at gmail.com (pravin.d.s at gmail.com) Date: Fri, 31 Jan 2014 13:44:08 +0530 Subject: [Lohit-devel-list] Lohit Malayalam test results In-Reply-To: <52EB373D.2080006@gmail.com> References: <52EB373D.2080006@gmail.com> Message-ID: On 31 January 2014 11:10, Santhosh Thottingal wrote: > Lohit Malayalam test results using SMC's test suite attached for > https://fedorahosted.org/releases/l/o/lohit/lohit- > malayalam-ttf-2.91.0.tar.gz > > The pdf shows the rendering mistakes that Ani can point out to > Sneha/Pravin. The chillu ??? is mapped to ? + ? + zwj. That is should be > removed and keep only ? + ? + zwj > Agree. Even i did not found any reference for same. Will remove it. > > Also remove the glyph for Samvruthokaram which is non standard > Yes, that is regression after fixing in earlier version. [1] > There is a duplicate glyph for U+0030 > Yes > > There are a few glyphs without any ligature rules. eg: glyph206, glyph209 > Yes, we missed rule for those. > > lllamlm_viramamlm.half_vamlm glyph looked like a hack. > This ligature was there in lohit 2.5.4 as well. This come with bugfix [3] Yes adding direct ligature does not looks good solution as our aim is to have compact font. Same time need to find some reference for same. > Is there any reason to include Devangari danda and double danda in this > font? > Unicode says: 0D64 " and 0D65 " are reserved and For viram punctuation, use the generic Indic 0964 and 0965. [2] So ideally we should put Malayalam script full stop there. Does Malayalam have specific glyph for full stop? If not then i think we can remove it from font. Thanks santhosh for your valuable feedback, We will report appropriate bugs in github for same and will remove in next release. I am cc'ing this email to lohit-devel to get more feedback on same. Regards, Pravin Satpute 1. https://bugzilla.redhat.com/show_bug.cgi?id=1013183 2. http://www.unicode.org/charts/PDF/U0D00.pdf 3. https://bugzilla.redhat.com/show_bug.cgi?id=523166 -------------- next part -------------- An HTML attachment was scrubbed... URL: From samjnaa at gmail.com Fri Jan 31 08:52:08 2014 From: samjnaa at gmail.com (Shriramana Sharma) Date: Fri, 31 Jan 2014 14:22:08 +0530 Subject: [Lohit-devel-list] Lohit Malayalam test results In-Reply-To: References: <52EB373D.2080006@gmail.com> Message-ID: On Fri, Jan 31, 2014 at 1:44 PM, pravin.d.s at gmail.com wrote: > > 0D64 " and 0D65 " are reserved and For viram > punctuation, use the generic Indic 0964 and 0965. [2] So ideally we should > put Malayalam script full stop there. > > Does Malayalam have specific glyph for full stop? If not then i think we > can remove it from font. Dandas may not be required for Malayalam language orthography but they are required for Sanskrit language orthography in Malayalam script. Hence all Indic fonts (including Tamil) should contain the danda characters. And AFAICS no Indic scripts have any other full stop distinct from the regular full stop. The regular full stop should not be removed in any Indic script font since people expect to use it. Recently I saw even Marathi books in Devanagari script that use it. -- Shriramana Sharma ???????????? ???????????? From snekore at gmail.com Fri Jan 31 12:29:44 2014 From: snekore at gmail.com (Sneha Kore) Date: Fri, 31 Jan 2014 17:59:44 +0530 Subject: [Lohit-devel-list] Lohit Malayalam test results In-Reply-To: References: <52EB373D.2080006@gmail.com> Message-ID: Hi all, Thanks to all for such a thorough evaluation & testing for Lohit Malayalam. The mail by Santhosh Thottingal was truely informative from an improvement perspective. On Fri, Jan 31, 2014On 31 January 2014 11:10, Santhosh Thottingal < > santhosh00 at gmail.com> wrote: > >> Lohit Malayalam test results using SMC's test suite attached for >> https://fedorahosted.org/releases/l/o/lohit/lohit-malayalam-ttf-2.91.0.tar.gz >> >> The pdf shows the rendering mistakes that Ani can point out to >> Sneha/Pravin. The chillu ??? is mapped to ? + ? + zwj. That is should be >> removed and keep only ? + ? + zwj >> > > Agree. Even i did not found any reference for same. Will remove it. > I do agree as well. I have removed the mapping to ? + ? + zwj . >> Also remove the glyph for Samvruthokaram which is non standard >> > > Yes, that is regression after fixing in earlier version. [1] > As this is non-standard, it has to get removed . so the sfd file has been modified to reflect the changes . > > >> There is a duplicate glyph for U+0030 >> > > Yes > Now, we had kept a single glyph i.e. one to which unicode value & unicode char gas been assigned . >> There are a few glyphs without any ligature rules. eg: glyph206, glyph209 >> > > Yes, we missed rule for those. > The Lookups has been updated with the rules for proper formation of the ligature . >> lllamlm_viramamlm.half_vamlm glyph looked like a hack. >> > > This ligature was there in lohit 2.5.4 as well. This come with bugfix [3] > Yes adding direct ligature does not looks good solution as our aim is to > have compact font. Same time need to find some reference for same. > Yes. This is the bit we are hanging in. yet not getting the solution for this. Better if you have a look at link http://www.unicode.org/L2/L2007/07393-mala-dot-reph.pdf (page 6) {specifically v2, v3 & v3 sections } . Please let us know if you have any solution. Will surely work in that direction. > > >> Is there any reason to include Devangari danda and double danda in this >> font? >> > Pravin Satpute & Shriramana Sharma had already answered about this. Ani Peter had helped in design issues. Shriramana Sharma had explained the thing about inclusion of dandas, very nicely. Thank you once again for the feedback. I had reported & resolved all these issues (except the one i.e. lllamlm_viramamlm.half_vamlm ) on github https://github.com/pravins/lohit/issues?state=open. The changes are now updated in lohit2. Those will be updated in lohit in next release as pravin has said. Thanks & Regards , Sneha Kore On Fri, Jan 31, 2014 at 2:22 PM, Shriramana Sharma wrote: > On Fri, Jan 31, 2014 at 1:44 PM, pravin.d.s at gmail.com > wrote: > > > > 0D64 " and 0D65 " are reserved and For viram > > punctuation, use the generic Indic 0964 and 0965. [2] So ideally we > should > > put Malayalam script full stop there. > > > > Does Malayalam have specific glyph for full stop? If not then i think we > > can remove it from font. > > Dandas may not be required for Malayalam language orthography but they > are required for Sanskrit language orthography in Malayalam script. > Hence all Indic fonts (including Tamil) should contain the danda > characters. > > And AFAICS no Indic scripts have any other full stop distinct from the > regular full stop. The regular full stop should not be removed in any > Indic script font since people expect to use it. Recently I saw even > Marathi books in Devanagari script that use it. > > -- > Shriramana Sharma ???????????? ???????????? > > _______________________________________________ > Lohit-devel-list mailing list > Lohit-devel-list at redhat.com > http://www.redhat.com/mailman/listinfo/lohit-devel-list -------------- next part -------------- An HTML attachment was scrubbed... URL: