-
Notifications
You must be signed in to change notification settings - Fork 346
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
GhostScript 9.27 released with yet another bug #880
Comments
How did you install ghostscript 9.27? The ghostscript from homebrew is still in 9.26. I also don't find a binary package from ghostscript website. |
I am using macports and they finally updated to 9.27 today. Surprised the homebrew is behind - they seem to be ahead of macports for most things. 9.27 was released 4 months ago. |
Given the rate of bugs they have had I decided to stay with 2.22 until I get confident to update again. Would have to build the ghost to try that and I’m a bit short of time for that these days.
From: Paul Wessel <[email protected]>
Sent: Tuesday, June 11, 2019 4:14 AM
To: GenericMappingTools/gmt <[email protected]>
Cc: Subscribed <[email protected]>
Subject: [GenericMappingTools/gmt] GhostScript 9.27 released with yet another bug (#880)
Fixed one bug, introduced another
I think they have fixed the #195<#195> but added another bug. The attached minimal pure PostScript example demonstrates the problem. It looks fine in previewers as a full US Letter page. To crop tightly we need the bounding box. In gs 9.26 we get the correct box:
gs -q -dSAFER -dNOPAUSE -dBATCH -sDEVICE=bbox bug.ps
which gives
%%BoundingBox: 71 71 361 385
%%HiResBoundingBox: 71.278803 71.278803 360.721184 384.173988
while in 9.27 we now get
%%BoundingBox: 71 71 361 377
%%HiResBoundingBox: 71.278803 71.278803 360.721184 376.541989
The smaller y-max means we crop across the Title and we get an error. Before I post this to the ghostscript issue tracker, can at least one confirm that this is the case please?
bug.ps.zip<https://github.com/GenericMappingTools/gmt/files/3274611/bug.ps.zip>
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub<#880?email_source=notifications&email_token=AAEDF2PRDBCI2SNBJZOCSQLPZ4J7PA5CNFSM4HW2FPP2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4GYWQTKA>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AAEDF2JLP6FJQGWDADBIYZDPZ4J7PANCNFSM4HW2FPPQ>.
|
I can reproduce the issue with the Linux x64 binary package provided by ghostscript website. |
Thanks, I have opened a bug report with ghostscript: https://bugs.ghostscript.com/show_bug.cgi?id=701186. |
Are there really no other viewers for Windows? Presumably you will be OK producing PDFs instead. |
Not an issue for me that have both gsview 5 and 6 but highly disturbing news. |
Well, you know the old story about rats and sinking ships. |
Yes, but there is also that story about the Ghost ships squad (Mary Celeste, of its name, was the head of them). Nice name for a gsview replacement. |
My bug report was seen as a duplicate of another report and apparently the problem has been fixed in the repo (https://bugs.ghostscript.com/show_bug.cgi?id=700952). |
Is there an alternative to postscript to create figures? Will this ever be necessary? It appears to me that postscript is becoming more and more peripheral, and the tools to show and process (at least free/open source) are fading away. Is this true? As of know, GMT is totally dependant on gs, right? Any danger that gs will disappear? What happens then? I do not know my gs history, so please dont keelhaul me. |
Nope. Long time concern of mine too. The closest I can imagine now would be postscript lib calls replaced by ones that would use Cairo. While some of them do not seem that different, there are many who are. This would be a large piece of work, though doable. I think. |
Homebrew finally updated to ghostscript 9.27 a few days ago. gs9.27 is known to be buggy, so I tried to install development version using Unfortunately, the gs development version has more serious bugs. I generated a PS file and tried to convert it to other formats:
Following command generated an empty PDF file.
Following command don't generate a JPG file.
|
What is the WARNING from psconvert about? |
Do you mean this warning?
I have the same warning with gs 9.26. |
Yes, given what it says I suspect something with the GeoPDF comments @joa-quim adds, but we can see what he things tomorrow. The message is not in the GMT source code. |
Try running the raw gs command reported by -S to see if we get more useful message from gs directly. |
The message comes from psconvert.c: Lines 2033 to 2040 in 53042bf
|
Yep, I grepped for too much and @joa-quim breaks lines to fit his tiny screen. |
Run psconvert with -S option:
Copy the raw gs command and run it give nothing. No errors, no test.jpg generated. Removing the |
Maybe see if removing any of the -d choices affects the outcome, i.e., find the fewest arguments that still fails so we can report a simple case. |
I removed all -d choices and used following command:
gs development version still fails to generate the JPG file and gives no warnings. However, gs 9.26 works well. |
Presumably the same with using test.ps directly? Please post that file (just in case) and gs --version output and I can submit a bug report. Alternatively, you could do the same at bugs.ghostscript.com. |
EPS file (zipped): psconvert_59910d.eps.zip gs command
gs is installed via |
I have submitted a bug report for master regarding this case, at https://bugs.ghostscript.com/show_bug.cgi?id=701272. |
Developers say: |
And it's only getting narrower. Because my eyes are f I'm forced to increase font size. Maybe we should simply shut up that warning. Things initiated long ago when in our (almost unused -J proj4 codes we have an -Jxy. But probably I don't know anymore what to do with it and when psconvert finds that for likely geog limits, it issues that warning. |
This gs (9.28) command doesn't work:
This gs (9.28) command works well:
The only difference is the |
Update: At least macports have moved to @9.27_1 from @9.27_0 and this fixes the original bug reported at the top. This is a patch since there is no 9.28 out, but it does mean that perhaps most non-Linux users will be OK by just getting the latest from fink, macports, home-brew. I just noticed that my gs was updated June 27 to 9.27_1. |
Hi, |
Yes, I think so. |
* Add affiliation and ORCID for Liam * Add Will Schlitzer ORCID * Add Dongdong Tian ORCID * Add Michael Grund ORCID * Clarify what should be added to AUTHORS.md * Add Malte Ziebarth to AUTHORS.md * Add Wei Ji ORCID and affiliation * Add Tyler Newton ORCHID & affiliation (GenericMappingTools#880) Co-authored-by: Dongdong Tian <[email protected]> Co-authored-by: Wei Ji <[email protected]> Co-authored-by: Meghan Jones <[email protected]> Co-authored-by: Malte Ziebarth <[email protected]> Co-authored-by: Will Schlitzer <[email protected]> Co-authored-by: Michael Grund <[email protected]> Co-authored-by: Tyler Newton <[email protected]>
Fixed one bug, introduced another
I think they have fixed the #195 but added another bug. The attached minimal pure PostScript example demonstrates the problem. It looks fine in previewers as a full US Letter page. To crop tightly we need the bounding box. In gs 9.26 we get the correct box:
gs -q -dSAFER -dNOPAUSE -dBATCH -sDEVICE=bbox bug.ps
which gives
while in 9.27 we now get
The smaller y-max means we crop across the Title and we get an error. Before I post this to the ghostscript issue tracker, can at least one confirm that this is the case please?
bug.ps.zip
The text was updated successfully, but these errors were encountered: