maxima-discuss — Maxima discussion list

Re: [Maxima-discuss] ***SPAM*** Re: Problems with the 5.48.0 tarball

[Viktor T. T. <vt...@vt...>]

Having seen this discussion, I also gave it a try.

I get:

(%i1) browser:"firefox '~A'";

(%o1)                            firefox '~A'
(%i2) output_format_for_help: html;

(%o2)                                html
(%i3) ? plot2d

and Firefox starts. I do NOT get the warning about mismatched info vs. 
html entries.

However, the browser invocation prefixes the path name with my local 
folder, so Firefox reports this:

Firefox can’t find the file at 
/home/vttoth/'file:///usr/local/share/maxima/5.48.0/doc/html/maxima_12.html\#index-plot2d`.

If I remove the "/home/vttoth/" bit and the quote-backquote, the link 
works. Oracle OS 8 (RHEL 8), texinfo 6.5.


Viktor




On 7/31/2025 8:44 PM, Jaime Villate via Maxima-discuss wrote:
> On 31/07/25 23:37, Raymond Toy wrote:
>> I think Xmaxima and wxMaxima are supposed to set it to frontend. If 
>> you manually set it, it doesn't work. I think that's because it 
>> doesn't then know where to send the output. I can't remember now. I 
>> should have at least left better comments in the code about this.
>
> First of all I'd like to understand what you would expect to happen 
> when the output is set to "frontend". Do you expect the frontend to go 
> and fetch the HTML and display it on its graphical interface?
>
> In that case the frontend should now how to parse HTML. Xmaxima is 
> able to parse some limited subset of HTML that does not include CSS or 
> javascript. It wouldn't do a good to display the Maxima manual. It is 
> better to let a real browser do the job. I doubt that Wxmaxima would 
> be able to parse HTML+CSS+Javascript either.
>
> Regards,
>
> Jaime
>
>
>
> _______________________________________________
> Maxima-discuss mailing list
> Max...@li...
> https://lists.sourceforge.net/lists/listinfo/maxima-discuss

Re: [Maxima-discuss] ***SPAM*** Re: Problems with the 5.48.0 tarball

[Jaime V. <vi...@fe...>]

On 31/07/25 23:37, Raymond Toy wrote:
> I think Xmaxima and wxMaxima are supposed to set it to frontend. If 
> you manually set it, it doesn't work. I think that's because it 
> doesn't then know where to send the output. I can't remember now. I 
> should have at least left better comments in the code about this.

First of all I'd like to understand what you would expect to happen when 
the output is set to "frontend". Do you expect the frontend to go and 
fetch the HTML and display it on its graphical interface?

In that case the frontend should now how to parse HTML. Xmaxima is able 
to parse some limited subset of HTML that does not include CSS or 
javascript. It wouldn't do a good to display the Maxima manual. It is 
better to let a real browser do the job. I doubt that Wxmaxima would be 
able to parse HTML+CSS+Javascript either.

Regards,

Jaime

Re: [Maxima-discuss] ***SPAM*** Re: Problems with the 5.48.0 tarball

[Jaime V. <vi...@fe...>]

On 31/07/25 23:49, Raymond Toy wrote:
> What is the value of |browser| and |url_base|? Alternatively, set 
> |*debug-display-html-help*| to |T|. I'm pretty sure this worked on 
> Windows before, but I don't know for sure because I rarely if ever use 
> Windows.

browser is set to "start ~A", just as the manual (that I guess you 
wrote) says. It seems wrong because it should include the name of the 
browser.

I then tried:

     browser: "start firefox ~A"
     ? plot2d

and the error message was:

   'start "firefox" "file:' is not recognized as an internal or external 
command.

To try to see what comes after file:, I then tried:

    browser: "start firefox ~S"
    ? plot2d

and the error message now shows me the whole URL:

    'start" "firefox" 
"\"file://C//maxima-current/share/maxima/branch_5_47_base_2097_gfbaf77047/doc/html/Plotting.html#index-plot2d\"' 
is not recognized as an internal or external command.

If I type in a terminal:

    start firefox 
file://C//maxima-current/share/maxima/branch_5_47_base_2097_gfbaf77047/doc/html/Plotting.html#index-plot2d

It works. The manual page for plot2d is shown in the browser.

There are two problems then. The default value of browser in Windows 
should include the name of a browser and there are some extra double 
quotes in the command passed to Windows.

  Regards,

Jaime

Re: [Maxima-discuss] ***SPAM*** Re: Problems with the 5.48.0 tarball

[Raymond T. <toy...@gm...>]

On 7/31/25 11:39 AM, Jaime Villate wrote:

> Here are the results of changing the default value of variable 
> "output_format_for_help" in Windows. There is no Maxima 5.48.0 in 
> Windows yet, but I'm testing with a very recent nightly buid of 
> 5.47.0-post in Wolfgang's site (it was built using texinfo 7.2).
>
> 1 - From command line:
>
> output_format_for_help: html  --> gives the warning about 2 HTML 
> entries missing
>
> ? plot2d --> fails with the following error message:
> 'start" "file:' is not recognized as an internal or external command, 
> operable program or batch file.

What is the value of |browser| and |url_base|? Alternatively, set 
|*debug-display-html-help*| to |T|. I'm pretty sure this worked on 
Windows before, but I don't know for sure because I rarely if ever use 
Windows.

I'm guessing that since the error message is "file:" is not recognized, 
|browser| is not set. It defaults to "firefox". Maybe on Windows, "edge" 
or whatever it's called is a better default.

>
>
> 3- Running from Wxmaxima:
>
> output_format_for_help: html  --> gives the warning about 2 HTML 
> entries missing
>
> ? plot2d --> fails with the same error message as in the command line:
> 'start" "file:' is not recognized as an internal or external command, 
> operable program or batch file.
>
> output_format_for_help: frontend --> it works fine
> ? plot2d -->
> <html-manual-keywords><keyword>plot2d</keyword></html-manual-keywords>
wxMaxima is supposed to set the value. I'm pretty sure this used to 
work, but I'm not a wxMaxima user. Maybe there's a menu or something 
that says to use html output? The fact that setting it to frontend 
doesn't produce an error tells me that wxMaxima redefined the function 
|display-frontend-topics| to accept frontend.
>
> No error, but the documentation for plot2d is not shown.
I'm not suprised. There needs to be coordination between wxMaxima to get 
this to work correctly. wxMaxima should have set |$maxima_frontend| to 
the appropriate thing and set |*help-display-function*| to 
|display-frontend-topics|, which should be defined by wxMaxima somewhere.
&#8203;

Re: [Maxima-discuss] ***SPAM*** Re: Problems with the 5.48.0 tarball

[Raymond T. <toy...@gm...>]

On 7/31/25 11:10 AM, Jaime Villate wrote:

> On 31/07/25 18:36, Jaime Villate via Maxima-discuss wrote:
>
> Here is the result, with Maxima 5.48.0 built with texinfo 6.5:
>
> (%i2) output_format_for_help: html;
> Warning: Number of HTML entries (3138) does not match text entries (3136)
>
> Warning: 2 HTML entries not in text entries:
>   "Error and warning messages" "Package zeilberger"
>
> (%o2)                                     html
> (%i3) ? plot2d;
> gio:file:///home/villate/.local/share/maxima/5.48.0/doc/html/Plotting.html#index-plot2d:
> Error when getting information for file “/home/villate/.local/share/maxima/5.48.0/doc/html/Plotting.html”:
>   No such file or directory
>
> (What's "gio"?)
>
> If I build Maxima 5.48.0 with texinfo 6.8, it now works, but it still 
> shows the same warning about the number of files:
>
> (%i2) output_format_for_help: html;
> Warning: Number of HTML entries (3138) does not match text entries (3136)
>
> Warning: 2 HTML entries not in text entries:
>   "Error and warning messages" "Package zeilberger"
Yeah, I haven't figured out why this is happening. And then I forgot 
about. Sorry about that. :-( This should really be fixed.
> (%o2)                                     html
> (%i3) ? plot2d;
> Opening in existing browser session.
>
> I'm running Maxima from Xmaxima, so let's try now the option "frontend":
>
> (%i4) output_format_for_help: frontend;
> (%o4)                                   frontend
> (%i5) ? plot2d;
> output_format_for_help: frontend not implemented.
>   -- an error. To debug this try: debugmode(true);
I think Xmaxima and wxMaxima are supposed to set it to frontend. If you 
manually set it, it doesn't work. I think that's because it doesn't then 
know where to send the output. I can't remember now. I should have at 
least left better comments in the code about this.
​

Re: [Maxima-discuss] Problems with the 5.48.0 tarball

[Viktor T. T. <vt...@vt...>]

Dear Jaime,

Hmmm. Why do we even have the generated .html files in dist-gzip I 
wonder? As you say, these are built from the tarball. So that means they 
should not be included in the tarball.


Viktor



On 7/31/2025 6:59 AM, Jaime Villate wrote:
> On 29/07/25 23:26, Viktor T. Toth wrote:
>> Dear Wolfgang,
>>
>> My suggestion: let's wait another day or two to see if there are any 
>> more significant issues and then I'll quickly do a 5.48.1 maintenance 
>> release; use that as the basis for an official Windows version.
>>
>> Let me ponder the idea of replacing make dist-gzip with using git 
>> directly. Let me check in particular if it plays nice with RPMs. (It 
>> should, but one never knows.) I kind of think though that dist-gzip 
>> should be preferred if it works; the build process should not require 
>> the presence of git.
>>
>>
>> Viktor
>
> Hello,
>
> I see two more significant issues:
>
> 1- If a user chooses one of the options --enable-lang-<code> of 
> configure, with <code> equal to de,  ja or ru, then make will fail 
> complaining the texi files in the doc/info/<code> directory are 
> missing. (thanks to James Cloos for telling me about it).
>
> We should either include those texi files in the tarball, or remove 
> options  --enable-lang-de, --enable-lang-ja and --enable-lang-ru from 
> the configure file, as well as removing directories doc/info/de, 
> doc/info/ja and doc/info/ru. In fact, all translations of the manual 
> (including es, pt and pt_BR) are very outdated and should probably be 
> removed from configure.
>
> 2- If I build the tarball with "make dist-gzip" the result is very 
> different from what Viktor gets. Since he is using makeinfo 6.5, a lot 
> of files maxima_<number>.html will be created. In my case, running 
> makeinfo 6.8, instead of those files I get <package_title>.html. That 
> means that if a user has texinfo 6.8 or later (as recommended by 
> configure), when building Maxima from the tarball distributed by 
> Viktor, he will end up with files maxima_<number>.html and 
> <package_title>.html which are two versions of the same manual sections.
>
> It can be argued that 2 is not a significant issue because Maxima 
> still installs and works fine, but the files duplication means more 
> than 100 additional files installed and an extra 21 MB, without any need.
>
> Regards,
>
> Jaime

Re: [Maxima-discuss] ***SPAM*** Re: Problems with the 5.48.0 tarball

[Jaime V. <vi...@fe...>]

Here are the results of changing the default value of variable 
"output_format_for_help" in Windows. There is no Maxima 5.48.0 in 
Windows yet, but I'm testing with a very recent nightly buid of 
5.47.0-post in Wolfgang's site (it was built using texinfo 7.2).

1 - From command line:

output_format_for_help: html  --> gives the warning about 2 HTML entries 
missing

? plot2d --> fails with the following error message:
'start" "file:' is not recognized as an internal or external command, 
operable program or batch file.

output_format_for_help: frontend --> After the warning for the 2 missing 
entries, it says:

output_format_for_help set to frontend, but no frontend is running.
  -- an error. To debug this try: debugmode(true);

Good, it kept the older value of output_format_for_help after the error 
message.


2- Running from Xmaxima:

output_format_for_help: html  --> gives the warning about 2 HTML entries 
missing

? plot2d --> fails with the same error message as in the command line:
'start" "file:' is not recognized as an internal or external command, 
operable program or batch file.

output_format_for_help: frontend --> it works fine
? plot2d ..>
output format for help: frontend not implemented.
  -- an error. To debug this try: debugmode(true);


3- Running from Wxmaxima:

output_format_for_help: html  --> gives the warning about 2 HTML entries 
missing

? plot2d --> fails with the same error message as in the command line:
'start" "file:' is not recognized as an internal or external command, 
operable program or batch file.

output_format_for_help: frontend --> it works fine
? plot2d -->
<html-manual-keywords><keyword>plot2d</keyword></html-manual-keywords>

No error, but the documentation for plot2d is not shown.

Jaime

Re: [Maxima-discuss] ***SPAM*** Re: Problems with the 5.48.0 tarball

[Jaime V. <vi...@fe...>]

On 31/07/25 18:36, Jaime Villate via Maxima-discuss wrote:
> Can you please remind us the setting we need to use to get help in 
> HTML format instead of info?

I'm answering myself:

-----------------------------------------------------------------------------------------------------------------

Option variable: output_format_for_help
Default value: text

output_format_for_help controls how describe displays help.

output_format_for_help can be set to one of the following values:

text
Help is displayed as plain text sent to a terminal. This is the default.

html
Help is displayed using a browser to display the HTML version of the manual.

frontend
Help is displayed using the frontend’s help system. If no frontend is 
running then an error is signaled. For example, wxMaxima and xmaxima are 
some frontends for maxima.

-----------------------------------------------------------------------------------------------------------------

Here is the result, with Maxima 5.48.0 built with texinfo 6.5:

(%i2) output_format_for_help: html;
Warning: Number of HTML entries (3138) does not match text entries (3136)

Warning: 2 HTML entries not in text entries:
  "Error and warning messages" "Package zeilberger"

(%o2)                                     html
(%i3) ? plot2d;
gio:file:///home/villate/.local/share/maxima/5.48.0/doc/html/Plotting.html#index-plot2d:
Error when getting information for file “/home/villate/.local/share/maxima/5.48.0/doc/html/Plotting.html”:
  No such file or directory

(What's "gio"?)

If I build Maxima 5.48.0 with texinfo 6.8, it now works, but it still 
shows the same warning about the number of files:

(%i2) output_format_for_help: html;
Warning: Number of HTML entries (3138) does not match text entries (3136)

Warning: 2 HTML entries not in text entries:
  "Error and warning messages" "Package zeilberger"

(%o2)                                     html
(%i3) ? plot2d;
Opening in existing browser session.

I'm running Maxima from Xmaxima, so let's try now the option "frontend":

(%i4) output_format_for_help: frontend;
(%o4)                                   frontend
(%i5) ? plot2d;
output_format_for_help: frontend not implemented.
  -- an error. To debug this try: debugmode(true);

So it does not work as advertised in the manual. I wonder what would 
happen in Wxmaxima (I don't have it in Ubuntu, but I will soon find out 
what happens in Wolfgang's nightly build for Windows).

Regards,

Jaime

Re: [Maxima-discuss] ***SPAM*** Re: Problems with the 5.48.0 tarball

[Jaime V. <vi...@fe...>]

On 31/07/25 15:58, Raymond Toy wrote:
> Oh. I haven't tried this but does the html help work with numbered 
> html files? If not, that would be unfortunate for the release.

Can you please remind us the setting we need to use to get help in HTML 
format instead of info?

Jaime

Re: [Maxima-discuss] ***SPAM*** Re: Why both guess.mac and devine.mac

[Jaime V. <vi...@fe...>]

On 31/07/25 18:22, Jaime Villate via Maxima-discuss wrote:
>
> You can see the message from where guess.mac and guess.texi came from, 
> in our mailing-lis archive:
>
> https://maxima.sourceforge.io/ext/list_archives/2002/msg01624.ht
>
The correct link is: 
https://maxima.sourceforge.io/ext/list_archives/2002/msg01624.html

Jaime

Re: [Maxima-discuss] ***SPAM*** Re: Why both guess.mac and devine.mac

[Jaime V. <vi...@fe...>]

On 31/07/25 15:24, Robert Dodier wrote:
> Huh, looks like I committed both of those, from work by Martin Rubey,
> posted to the mailing list.
No problem. Nobody expects you to remember what you committed 18 years 
earlier :)
>   They're not quite the same. I fixed a
> couple of trivial bugs in the older copy (devine.mac). However,
> guess.mac is from a later message from Martin which is titled "bug
> fixes" or something like that. So there are changes in both versions.

You can see the message from where guess.mac and guess.texi came from, 
in our mailing-lis archive:

https://maxima.sourceforge.io/ext/list_archives/2002/msg01624.ht

> I think I didn't notice that guess.mac and devine.mac are different
> versions of the same thing. At this point it seems like the course of
> action is to merge the bug fixes I made to devine.mac into guess.mac,
> delete the former, and rename the latter to a more descriptive name as
> suggested.

That sounds like a good plan.

Jaime

Re: [Maxima-discuss] Announcing Maxima 5.48

[<pto...@gm...>]

a corrected 5.48.1 release should occur soon, for which I will build a Windows installer.
[PT] Thanks for the update and explanation.  I'll keep an eye out for version 5.48.1 for Windows.

--Pete


Best regards, Wolfgang

Re: [Maxima-discuss] Problems with the 5.48.0 tarball

[Raymond T. <toy...@gm...>]

On 7/31/25 3:59 AM, Jaime Villate via Maxima-discuss wrote:

> On 29/07/25 23:26, Viktor T. Toth wrote:
>
>
> 2- If I build the tarball with "make dist-gzip" the result is very 
> different from what Viktor gets. Since he is using makeinfo 6.5, a lot 
> of files maxima_<number>.html will be created. In my case, running 
> makeinfo 6.8, instead of those files I get <package_title>.html. That 
> means that if a user has texinfo 6.8 or later (as recommended by 
> configure), when building Maxima from the tarball distributed by 
> Viktor, he will end up with files maxima_<number>.html and 
> <package_title>.html which are two versions of the same manual sections.
Oh. I haven't tried this but does the html help work with numbered html 
files? If not, that would be unfortunate for the release.
&#8203;

Re: [Maxima-discuss] Why both guess.mac and devine.mac

[Stavros M. <mac...@gm...>]

True that they're not quite the same, but comparing them takes more than a
glance, because variables have been renamed....

Thanks for working on this!

The documentation in guess/guess.info is very helpful. There is one thing
that would be useful to clarify. It reads "of the form <rational function>,
<product of rational function>, <product of product of rational function>,"
where "product" doesn't mean just multiplication, but an indexed product
like product(..., i, ...). That explains why it handles 2^i, for example.

The use of the strings "one" and "nogamma" aren't really idiomatic in
Maxima language. I'd think they should be symbols.

Guess sometimes returns subexpressions like abs(i1+1). Perhaps the
i*n's* should
be declared non-negative integer.



On Thu, Jul 31, 2025 at 10:24 AM Robert Dodier <rob...@gm...>
wrote:

> On Wed, Jul 30, 2025 at 2:43 PM Stavros Macrakis <mac...@gm...>
> wrote:
>
> > In 5.48, we have both 5.48.0/share/contrib/devine.mac and
> 5.48.0/share/contrib/guess/*, which have identical executable code, but
> guess/* includes documentation.
> >
> > Shouldn't devine.mac be deleted?
>
> Huh, looks like I committed both of those, from work by Martin Rubey,
> posted to the mailing list. They're not quite the same. I fixed a
> couple of trivial bugs in the older copy (devine.mac). However,
> guess.mac is from a later message from Martin which is titled "bug
> fixes" or something like that. So there are changes in both versions.
>
> I think I didn't notice that guess.mac and devine.mac are different
> versions of the same thing. At this point it seems like the course of
> action is to merge the bug fixes I made to devine.mac into guess.mac,
> delete the former, and rename the latter to a more descriptive name as
> suggested.
>
> FWIW
>
> Robert
>

Re: [Maxima-discuss] Why both guess.mac and devine.mac

[Robert D. <rob...@gm...>]

On Wed, Jul 30, 2025 at 2:43 PM Stavros Macrakis <mac...@gm...> wrote:

> In 5.48, we have both 5.48.0/share/contrib/devine.mac and 5.48.0/share/contrib/guess/*, which have identical executable code, but guess/* includes documentation.
>
> Shouldn't devine.mac be deleted?

Huh, looks like I committed both of those, from work by Martin Rubey,
posted to the mailing list. They're not quite the same. I fixed a
couple of trivial bugs in the older copy (devine.mac). However,
guess.mac is from a later message from Martin which is titled "bug
fixes" or something like that. So there are changes in both versions.

I think I didn't notice that guess.mac and devine.mac are different
versions of the same thing. At this point it seems like the course of
action is to merge the bug fixes I made to devine.mac into guess.mac,
delete the former, and rename the latter to a more descriptive name as
suggested.

FWIW

Robert

Re: [Maxima-discuss] Problems with the 5.48.0 tarball

[Jaime V. <vi...@fe...>]

On 29/07/25 23:26, Viktor T. Toth wrote:
> Dear Wolfgang,
>
> My suggestion: let's wait another day or two to see if there are any 
> more significant issues and then I'll quickly do a 5.48.1 maintenance 
> release; use that as the basis for an official Windows version.
>
> Let me ponder the idea of replacing make dist-gzip with using git 
> directly. Let me check in particular if it plays nice with RPMs. (It 
> should, but one never knows.) I kind of think though that dist-gzip 
> should be preferred if it works; the build process should not require 
> the presence of git.
>
>
> Viktor

Hello,

I see two more significant issues:

1- If a user chooses one of the options --enable-lang-<code> of 
configure, with <code> equal to de,  ja or ru, then make will fail 
complaining the texi files in the doc/info/<code> directory are missing. 
(thanks to James Cloos for telling me about it).

We should either include those texi files in the tarball, or remove 
options  --enable-lang-de, --enable-lang-ja and --enable-lang-ru from 
the configure file, as well as removing directories doc/info/de, 
doc/info/ja and doc/info/ru. In fact, all translations of the manual 
(including es, pt and pt_BR) are very outdated and should probably be 
removed from configure.

2- If I build the tarball with "make dist-gzip" the result is very 
different from what Viktor gets. Since he is using makeinfo 6.5, a lot 
of files maxima_<number>.html will be created. In my case, running 
makeinfo 6.8, instead of those files I get <package_title>.html. That 
means that if a user has texinfo 6.8 or later (as recommended by 
configure), when building Maxima from the tarball distributed by Viktor, 
he will end up with files maxima_<number>.html and <package_title>.html 
which are two versions of the same manual sections.

It can be argued that 2 is not a significant issue because Maxima still 
installs and works fine, but the files duplication means more than 100 
additional files installed and an extra 21 MB, without any need.

Regards,

Jaime

Re: [Maxima-discuss] Announcing Maxima 5.48

[Wolfgang D. <wol...@da...>]

Am 31.07.25 um 04:09 schrieb pto...@gm...:
> I went to the link given in the note below and tried to find the download for Windows.  One of the links on that page went to:
> https://sourceforge.net/projects/maxima/files/
> where the latest version of Maxima for Windows seems to be 5.47.  Is there a version 5.48 for Windows or is 5.47 the latest?

Hi Pete,

I usually compile the Windows installers.
The 5.48 release is missing some files, which prevent the automatic 
build of a Windows installer too, see this discussion:

https://sourceforge.net/p/maxima/mailman/maxima-discuss/thread/ec8ee766-c13d-42c2-a1e3-7717354bc5fd%40fe.up.pt/

I asked, if I should build it anyway (by adding the missing files), but 
a corrected 5.48.1 release should occur soon, for which I will build a 
Windows installer.

Best regards, Wolfgang

Re: [Maxima-discuss] Announcing Maxima 5.48

[<pto...@gm...>]

I went to the link given in the note below and tried to find the download for Windows.  One of the links on that page went to:
https://sourceforge.net/projects/maxima/files/
where the latest version of Maxima for Windows seems to be 5.47.  Is there a version 5.48 for Windows or is 5.47 the latest?

Thanks.

--Pete

-----Original Message-----
From: Viktor T. Toth <vt...@vt...

-----Original Message-----
From: Viktor T. Toth <vt...@vt...> 
Sent: Tuesday, July 29, 2025 12:47 AM
To: max...@li...
Subject: [Maxima-discuss] Announcing Maxima 5.48

Maxima version 5.48 has just been released.

The new version introduces several noteworthy enhancements for symbolic computation, with improvements in performance, functionality, and user experience.

🔍 Highlights:

- Unicode-enabled output (when supported by the Lisp compiler)
- Numerous performance improvements across core routines
- New package for symbolic radical denesting
- New package for inferring closed-form expressions from sequences
- New package for simplification of gamma functions
- Resolution of more than 150 tickets, spanning both long-standing and recent bugs

Developed in Common Lisp, Maxima remains a reliable and customizable tool for research, education, science, and engineering.

🔗 To install, explore, or contribute: https://maxima.sourceforge.io



_______________________________________________
Maxima-discuss mailing list
Max...@li...
https://lists.sourceforge.net/lists/listinfo/maxima-discuss

Re: [Maxima-discuss] Why both guess.mac and devine.mac

[Raymond T. <toy...@gm...>]

On 7/30/25 2:41 PM, Stavros Macrakis wrote:

> In 5.48, we have both 5.48.0/share/contrib/devine.mac 
> and 5.48.0/share/contrib/guess/*, which have identical executable 
> code, but guess/* includes documentation.
>
> Shouldn't devine.mac be deleted?
>
> The guess/* doc clarifies that in the text "expressions of the form 
> <rational function>, <product of rational function>, <product of 
> product of rational function>," "product" means an indexed finite 
> product like product(..., i, 1, n).
>
> I still think the function should be named something more explicit 
> (guess_rational_function_sequence?). cf. 
> https://sourceforge.net/p/maxima/bugs/4589/

Can't say much about the packages themselves, but "guess" really is too 
broad. "guess_rational_function_sequence" is much better, even if it is 
a lot to type.

&#8203;

[Maxima-discuss] Why both guess.mac and devine.mac

[Stavros M. <mac...@gm...>]

In 5.48, we have both 5.48.0/share/contrib/devine.mac
and 5.48.0/share/contrib/guess/*, which have identical executable code, but
guess/* includes documentation.

Shouldn't devine.mac be deleted?

The guess/* doc clarifies that in the text "expressions of the form
<rational function>, <product of rational function>, <product of product of
rational function>," "product" means an indexed finite product like
product(..., i, 1, n).

I still think the function should be named something more explicit
(guess_rational_function_sequence?). cf.
https://sourceforge.net/p/maxima/bugs/4589/

           -s

Re: [Maxima-discuss] core was built for runtime "hostname-id-1753844457" but this is "hostname-id-1753845453"

[eric <er...@co...>]

On 7/30/25 03:03, Nasser M. Abbasi wrote:
> I am on CachyOS linux (Arch based). Just did system update (pacman -Syu) 
> and saw it updated maxima to 5.48.
> 
> But when I start it, it gives error. Here is the info from pacman 
> package manager:
> 
>  >sudo pacman -Qi maxima
> Installed From  : extra
> Name            : maxima
> Version         : 5.48.0-3
> Description     : A sophisticated computer algebra system
> Architecture    : x86_64
> URL             : http://maxima.sourceforge.net
> Licenses        : GPL-2.0-only
> Groups          : None
> Provides        : None
> Depends On      : maxima-backend  texinfo
> Optional Deps   : gnuplot: plotting capabilities
>                    rlwrap: readline support via /usr/bin/rmaxima
>                    tk: graphical xmaxima interface [installed]
> Required By     : maxima-fas
> Optional For    : None
> Conflicts With  : None
> Replaces        : maxima-ecl<5.45.1-9
> Installed Size  : 77.35 MiB
> Packager        : Antonio Rojas <ar...@ar...>
> Build Date      : Wed 30 Jul 2025 01:15:33 AM CDT
> Install Date    : Wed 30 Jul 2025 04:47:12 AM CDT
> Install Reason  : Installed as a dependency for another package
> Install Script  : No
> Validated By    : Signature
> 
> 
> Now I do:
> 
>  >maxima
> fatal error encountered in SBCL pid 4350 tid 4350:
> core was built for runtime "hostname-id-1753844457" but this is 
> "hostname-id-1753845453"
> 
>  >which maxima
> /usr/bin/maxima
> 
>  >ls -lrt `which maxima`
> -rwxr-xr-x 1 root root 9474 Jul 30 01:15 /usr/bin/maxima
>  >
> 
> This could be arch linux packging issue? I do not know. Never see such 
> an error before. Reboted my Linux PC, still this error shows up.
> 
> Just thought to mention it in case someone knows what it could mean.
> 
> Regards
> --Nasser
> 

Maxima 5.48 is working on Arch linux. It was updated during a reqular 
system update a couple of days ago.

Maxima 5.48.0 https://maxima.sourceforge.io
using Lisp ECL 24.5.10

I don't know if that info will help narrow down the error you are having.

Regards,
Eric

Re: [Maxima-discuss] On Windows, rtest_draw, 179 tests failed out of 181 total tests

[Jaime V. <vi...@fe...>]

On 30/07/25 16:08, Jinsong Zhao wrote:
> According to 
> https://github.com/Clozure/ccl/issues/538#issuecomment-3136540176, the 
> behavior of SBCL and CCL is correct. So the definition in the 
> operatingsystem package should be revised.

Let's hope David Scherfgen has some time to look at operatingsystem.lisp 
and replace that SBCL specific part by some portable CL code :)

Jaime

Re: [Maxima-discuss] core was built for runtime "hostname-id-1753844457" but this is "hostname-id-1753845453"

[Robert D. <rob...@gm...>]

I don't know about this specific issue with Arch Linux, but I have
bumped into a similar error with Maxima + SBCL in other contexts. I
believe it's triggered by compiling Maxima with one version of SBCL,
creating a core dump, and trying to load that core dump into a
different version of SBCL.

A workaround is to create an executable instead of a core dump when
compiling. There is an option for configure to choose either a core
dump or an executable, I forget what it is exactly but configure
--help should show it.

People creating packages for distribution should probably configure
the build to create an executable instead of a core dump (whether for
SBCL or any other Lisp) in order to obviate differences between
versions.

Hope this helps,

Robert

Re: [Maxima-discuss] core was built for runtime "hostname-id-1753844457" but this is "hostname-id-1753845453"

[Nasser M. A. <nm...@12...>]

FYI;

Saw reply at sagemath forum after I posted my question there that this 
was an Arch Linux packging issue with SBCL and should be fixed.

Here is the message

<https://groups.google.com/g/sage-devel/c/WZjKcAvK2mY>

I will try to do upgrade of SBCL on my PC later on today and verify.

--Nasser


On 7/30/25 11:17 AM, Nasser M. Abbasi wrote:
> 
> 
> On 7/30/25 10:59 AM, Jaime Villate wrote:
>> On 30/07/25 11:03, Nasser M. Abbasi wrote:
>>> Now I do:
>>>
>>> >maxima
>>> fatal error encountered in SBCL pid 4350 tid 4350:
>>> core was built for runtime "hostname-id-1753844457" but this is 
>>> "hostname-id-1753845453"
>>>
>>> >which maxima
>>> /usr/bin/maxima
>>>
>>> >ls -lrt `which maxima`
>>> -rwxr-xr-x 1 root root 9474 Jul 30 01:15 /usr/bin/maxima
>>> >
>>>
>>> This could be arch linux packging issue? I do not know. Never see 
>>> such an error before. Reboted my Linux PC, still this error shows up.
>>
>> Hello,
>>
>> thanks for your message. What version of SBCL is being used. This 
>> looks similar to a problem we've had in Ubuntu. We could not build 
>> Maxima with SBCL anymore until we updated SBCL to a newer version.
>>
>> Regards,
>>
>> Jaime
>>
> 
> pacman says
> 
>  >pacman -Qi sbcl
> Installed From  : cachyos-extra-znver4
> Name            : sbcl
> Version         : 2.5.7-1.1
> Description     : Steel Bank Common Lisp
> Architecture    : x86_64_v4
> URL             : http://www.sbcl.org/
> Licenses        : BSD-2-Clause  BSD-3-Clause  Xerox  LOOP
> Groups          : None
> Provides        : common-lisp  cl-asdf
> Depends On      : zlib  zstd
> Optional Deps   : None
> Required By     : maxima-sbcl
> Optional For    : None
> Conflicts With  : None
> Replaces        : None
> Installed Size  : 71.87 MiB
> Packager        : CachyOS <ad...@ca...>
> Build Date      : Tue 29 Jul 2025 10:17:33 PM CDT
> Install Date    : Wed 30 Jul 2025 04:27:29 AM CDT
> Install Reason  : Installed as a dependency for another package
> Install Script  : No
> Validated By    : Signature
> 
> I also get this error when calling maxima.version() from inside sagemath
> 
> 
>  >sage
> ┌────────────────────────────────────────────────────────────────────┐
> │ SageMath version 10.6, Release Date: 2025-03-31                    │
> │ Using Python 3.13.5. Type "help()" for help.                       │
> └────────────────────────────────────────────────────────────────────┘
> sage: var('x')
> x
> sage: integrate(sin(x),x,algorithm="maxima")
> -cos(x)
> sage: print(maxima.version())
> fatal error encountered in SBCL pid 97224 tid 97224:
> core was built for runtime "hostname-id-1753844457" but this is 
> "hostname-id-1753845453"
> 
> ---------------------------------------------------------------------------
> IndexError                                Traceback (most recent call last)
> Cell In[3], line 1
> ----> 1 print(maxima.version())
> 
> File /usr/lib/python3.13/site-packages/sage/interfaces/ 
> maxima_abstract.py:431, in MaximaAbstract.version(self)
>      420 def version(self):
>      421     r"""
>      422     Return the version of Maxima that Sage includes.
>      423
>     (...)    429         '5.41.0'
>      430     """
> --> 431     return maxima_version()
> 
> File /usr/lib/python3.13/site-packages/sage/interfaces/ 
> maxima_abstract.py:2154, in maxima_version()
>     2142 """
>     2143 Return Maxima version.
>     2144
>     (...)   2151     '5.41.0'
>     2152 """
>     2153 with os.popen('{} --version'.format(MAXIMA)) as p:
> -> 2154     return p.read().split()[-1]
> 
> IndexError: list index out of range
> sage:
> 
> 
> So I do not know now if sagemath integrate command above is using maxima 
> 5.48 or sagemath buildin maxima.  I should ask sagemath folks on this.
> 
> I am on CachyOS Linux
> 
>  >uname -a
> Linux cachyos-x8664 6.16.0-2-cachyos #1 SMP PREEMPT_DYNAMIC Tue, 29 Jul 
> 2025 07:46:30 +0000 x86_64 GNU/Linux
>  >cat /etc/os-release
> NAME="CachyOS Linux"
> PRETTY_NAME="CachyOS"
> ID=cachyos
> BUILD_ID=rolling
> ANSI_COLOR="38;2;23;147;209"
> HOME_URL="https://cachyos.org/"
> DOCUMENTATION_URL="https://wiki.cachyos.org/"
> SUPPORT_URL="https://discuss.cachyos.org/"
> BUG_REPORT_URL="https://github.com/cachyos"
> PRIVACY_POLICY_URL="https://terms.archlinux.org/docs/privacy-policy/"
> LOGO=cachyos
>  >
> 
> 
> Any more info, please let me know.
> 
> Regards
> --Nasser
> 

Re: [Maxima-discuss] On Windows, rtest_draw, 179 tests failed out of 181 total tests

[David S. <d.s...@go...>]

Why is there even SBCL-specific code in the first place? This should be
possible in a portable way for all Lisps using the CL functions
ensure-directories-exist and merge-pathnames.

Jinsong Zhao <js...@ye...> schrieb am Mi., 30. Juli 2025, 02:31:

>
> On 7/30/2025 4:53 PM, Jaime Villate wrote:
> > Hello Jingsong,
> >
> > thank you for working on this.
> >
> > On 30/07/25 06:16, Jinsong Zhao wrote:
> >> In package operatingsystem.lisp, for sbcl the os-mkdir is defined as:
> >>
> >> #+sbcl (sb-unix:unix-mkdir (directory-namestring
> >> (sb-ext:native-pathname dir)) #o777)
> >>
> >> The problem is caused by directory-namestring, which discard the disk
> >> label in the pathname:
> >>
> >> * (directory-namestring "R:/Temp/abc/")
> >> "/Temp/abc/"
> >>
> >> I don't know what's the correct behavior of directory-namestring.
> >> However, in sbcl 2.2.9 on Debian and 2.5.6 on FreeBSD, both give:
> >> * (directory-namestring "R:/Temp/abc/")
> >> "R:/Temp/abc/"
> >>
> > Do you mean that (directory-namestring "R:/Temp/abc/") gives
> > "/Temp/abc/"when run in Windows, while it results in "R:/Temp/abc/"
> > when run in Debian and FreeBSD?
>
> Yes. However, I don't know which is the right answer. I just reported
> this issue to SBCL mailing list.
>
> I also test it on CCL in Windows, it gives "/Temp/abc/", however, CCL in
> FreeBSD gives "R\\:/Temp/abc/".  Maybe, I should also file a issue.
>
> >
> > directory-namestring seems to be important in cases like:
> >
> > * (directory-namestring "~/.maxima/maxima-init.mac")
> >
> > where you want to extract the directory name from the file path.
> >
> > Regards,
> >
> > Jaime
>
> Best,
>
> Jinsong
>
>
>
> _______________________________________________
> Maxima-discuss mailing list
> Max...@li...
> https://lists.sourceforge.net/lists/listinfo/maxima-discuss
>