Bug 178411
| Summary: | Cannot see Japanese dvi file because of vfontmap description | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Mamoru TASAKA <mtasaka> |
| Component: | tetex | Assignee: | Jindrich Novy <jnovy> |
| Status: | CLOSED RAWHIDE | QA Contact: | David Lawrence <dkl> |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | rawhide | CC: | eng-i18n-bugs, furutaka, pknirsch, tagoh, t.matsuu |
| Target Milestone: | --- | Keywords: | i18n, Reopened |
| Target Release: | --- | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2006-04-28 04:45:00 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
| Attachments: | |||
|
Description
Mamoru TASAKA
2006-01-20 09:22:00 UTC
Created attachment 123471 [details]
sample Japanese tex file ( in ISO-2022-JP )
Sample Japanese tex file.
It is encoded with ISO-2022-JP.
Created attachment 123473 [details]
Screenshot of pxdvi
Could you please confirm this is correct output when viewing the sample file
you attached?
(In reply to comment #2) > Created an attachment (id=123473) [edit] > Screenshot of pxdvi > > Could you please confirm this is correct output when viewing the sample file > you attached? Okay. The view of the attachment in the comment 2 is CORRECT. It looks like the rawhide fonts-japanese misses the kochi-mincho-subst font completely. When I install the fonts manualy from the old fonts-japanese, I'm able to run pxdvi smoothly, but the japanese fonts aren't correctly displayed. Comparing the output from pxdvi on my working FC4 and rawhide where it doesn't work, I see that this warning is displayed prior the other messages: Warning: Cannot convert string "-*-helvetica-medium-r-*-*-12-*-*-*-*-*-*" to type FontStruct I'm trying to figure out what's happening here (suspecting freetype). I tried to comment out the kochi fonts and substitute them with the sazanami ones in /usr/share/texmf/pxdvi/vfontmap. But with no luck... pxdvi displays the DVI contents, but japanese characters are displayed like they're missing in the font (black squares)... Yes. I think that changing description in vfontmap from kochi to sazanami is needed, but I have no idea for further fix for now... Akira, could you please help us here why the fonts are displayed incorrectly when vfontmap points on the sazanami fonts? Were there some important changes we are not aware of in fonts-japanese since fonts-japanese-0.20050222-3 that could affect pxdvi? Thanks. No idea off hand. Kochi fonts and Sazanami fonts are different TrueType font - properly which was made by even the same upstream though - I have also no idea how pxdvi picks the glyphs up from TrueType fonts. so need to investigate this anyway. let me have a look then. I just tried this out on Rawhide with kochi-{gothic,mincho}-subst.ttf that
copied manually. but it doesn't work for me. what you did try at Comment #2 was
on FC4 box, right? and what you were saying at Comment #4 was on Rawhide box?
I'm wondering if this is related to fonts. guessing from asking me about font,
you are sure there aren't any changes in pxdvi between FC4 and Rawhide right. so
I suspect it may be freetype or something like rendering library as you did then.
Ok, I've also tried this on FC4 box with sazanami-{gothic,mincho}.ttf. it works
fine. so this is apparently rendering issue.
To answer your question in Comment #9, yes, pxdvi worked fine under FC4, the output from Comment #4 was taken from an updated FC4. In Comment #4 is used the latest rawhide fonts-japanese on rawhide box, right. I applied change to vfontmap suggested by MATSUURA Takanori. Maybe that will fix your issue. Hello, Jindrich. Which version of rpm(s) do you refer to? I think the current newest rawhide tetex rpm is tetex-3.0-17 (in 20060307 rawhide) and I don't see any change in vfontmap... If your change will be applied in 20060310 rawhide, I will wait until then. Hello Mamoru, the latest tetex rpm version is 3.0-18, where the attempt to fix it is applied. Maybe it's not on the mirrors yet because of the FC5 freeze, but I built it few days ago. In the meantime the srpm is here: http://people.redhat.com/jnovy/files/tetex-3.0-18.src.rpm I just rebuilt rpms from source tetex-3.0-18 and installed them, but it seems to no purpose... The change in 3.0-18 is that vfontmap points to sazanami fonts, right? Then the attempt is already done as I reported on the Note of the first bug report and the comments from 1 to 6 says. Can you see the dvi file correctly (as the attachment in comment 2) created from the original tex file (attached in the comment 1) by 3.0-18 ? Note: when I create ps file from dvi file by pdvips -Ppdf sample.dvi , the created ps file can be seen correctly by gs, so I suppose there is no problem in dvi file itself. Created attachment 127061 [details]
A patch to change some definition in xdvi.h
Hello, Jindrich.
Current rawhide rpm tetex-3.0-19 does not work well, however, there is some
news.
A information that "Making it valid the definition that "typedef
unsigned int wide_ubyte;" in texk/pxdvik/xdvi.h should work for this issue"
was brought on Japanese Fedora wiki page, and this worked WELL for me!!
Currently, I apply the attached patch for tetex.spec and it works well for now.
I checked that on FC4 system NeedFunctionPrototypes was defined as 1 at the
line 234 of texk/pxdvik/xdvi.h, but on FC5 system NeedFunctionPrototypes is
not defined at this line, and by the following lines NeedFunctionPrototypes is
defined as 0. So I suppose that this is related to switching to modular X.
Hi Mamoru, yes, modular X seems to be at fault here as far as I can see from X11/Xfuncproto.h, which is included by X11/Xlib.h in xdvi.h: <snip> #ifndef NARROWPROTO #define NARROWPROTO #endif #ifndef FUNCPROTO #define FUNCPROTO 15 #endif #ifndef NeedWidePrototypes #ifdef NARROWPROTO #define NeedWidePrototypes 0 #else #define NeedWidePrototypes 1 /* default to make interropt. easier */ #endif #endif /* NeedWidePrototypes */ </snip> what causes that NeedWidePrototypes is defined to 0 if previously undefined. Mike, is it intentional? From the comment above in X11/Xfuncproto.h I'd guess the default should be #define NeedWidePrototypes 1 ? Ok, I applied the workaround presented by Mamoru. Please reopen if you see any other problems. The workaround is applied since tetex-3.0-20. *** Bug 187893 has been marked as a duplicate of this bug. *** Hello, Jindrich. tetex-3.0-20 works WELL for me, thanks. Then would you have a plan to apply these temporary fix (vfontmap change in 3.0-18 and definition of NeedWidePrototypes in 3.0-20) to FC5 updates ? Hi Mamoru, I modified the vfontmap file once more based on recommendation from Matsuura Takanori. This should appear soon including more japanese fixes since 3.0-22. Could you please confirm everything works as expected with this version? If so, I'll release updates immediatelly. Okay, I will check tetex-3.0-22 when it appears on web site. Hello, Jindrich. Now I checked tetex-3.0-22 and it seems to work well. tetex-3.0-19.fc5 has been pushed for fc5, which should resolve this issue. If these problems are still present in this version, then please make note of it in this bug report. Okay, I upgraded to tetex-3.0-19.fc5 on FC5 system and it seems to work well. Thanks!! Now, I close this bug as CLOSED RAWHIDE. If another bug is found, I will reopen this bug or create another bug report. Created attachment 128344 [details]
some warning on the background after pxdvi to view the file
I tried with tetex-3.0-22, and it showing fine too. Noticed some warning output
in the background thought. Please let me know if it should be of concern.
Created attachment 128463 [details] updated tetex-3.0-pxdvi.patch (In reply to comment #26) > some warning on the background after pxdvi to view the file This patch may fix the last "Warning: Cannot convert string..." warning. "Warning: Missing charsets in String to FontSet conversion" warnings are from font detection mechanism in pxdvi GUI and harmless for the pxdvi functions . xdvi command also output the same warning in my FC5. Other warnings are not shown in my FC5. Does your machine have fonts-japanese package required by tetex-xdvi? Could you open the new bug If more fix is needed? RE Comment #27 Sure. Thanks for your patch. tetex-3.0-10.FC4 has been pushed for fc4, which should resolve this issue. If these problems are still present in this version, then please make note of it in this bug report. tetex-3.0-10.FC4 has been pushed for fc4, which should resolve this issue. If these problems are still present in this version, then please make note of it in this bug report. |