RFC: path relative to current file.
Hello, list! A while ago, I raised a question about how to, in a tex, specify a file with a path relative to the currently being processed file. http://www.ntg.nl/pipermail/ntg-context/2012/069132.html http://www.ntg.nl/pipermail/ntg-context/2012/068902.html Even though the issue did not get much attention, I do believe it is the right thing to do. I wrote a module to implement a macro I called "\pathrelativetome". My code is still a lot ugly. I don't know how to do many things in Lua, LuaTeX and ConTeXt. It can be found here: https://bitbucket.org/andrecaldas/math-video-classes/src/bade74046a6b/lib It would be REALLY NICE if the functionality of "pathrelativetome" was provided by ConTeXt. call it using, for example: context --usemodule=lib/relativetome src/products/assorted.tex Ugly things remaining: 1. pathstack should be initialized with the path of the first tex script. I.e.: src/products/, instead of pathstack[1] = 'src/products' 2. Functions pop and push should be local. (relativetome.lua) 3. Call function "pop" automatically, just like the "push" function is called automatically. 4. Avoid having to define "\xproject", "\xproduct", "\xcomponent" and "\xenvironment". http://www.ntg.nl/pipermail/ntg-context/2012/069305.html Any suggestions/improvements are very welcome! Cheers, André Caldas.
On 13-10-2012 03:34, Andre Caldas wrote:
3. Call function "pop" automatically, just like the "push" function is called automatically.
kind of tricky as the file can already be loaded and closed and therefore you're back at the previous level
4. Avoid having to define "\xproject", "\xproduct", "\xcomponent" and "\xenvironment". http://www.ntg.nl/pipermail/ntg-context/2012/069305.html
just use \environment [somefilename] as a space delimited variant does not work well in \environment \somefilename ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
On 13-10-2012 03:34, Andre Caldas wrote:
Hello, list!
A while ago, I raised a question about how to, in a tex, specify a file with a path relative to the currently being processed file. http://www.ntg.nl/pipermail/ntg-context/2012/069132.html http://www.ntg.nl/pipermail/ntg-context/2012/068902.html
Even though the issue did not get much attention, I do believe it is the right thing to do. I wrote a module to implement a macro I called "\pathrelativetome". My code is still a lot ugly. I don't know how to do many things in Lua, LuaTeX and ConTeXt. It can be found here: https://bitbucket.org/andrecaldas/math-video-classes/src/bade74046a6b/lib
It would be REALLY NICE if the functionality of "pathrelativetome" was provided by ConTeXt.
in file-job.lua add around the function 'process': local function toppath() local pathname = dirname(inputstack[#inputstack] or "") if pathname == "" then return "." else return pathname end end resolvers.toppath = topath resolvers.prefixes.toppath = function(str) local fullname = cleanpath(joinpath(toppath(),str)) return fullname end local function process(what,name) local depth = #typestack local process -- name = resolvers.resolve(name) -- ..... keep code ... and somewhere at the top: local joinpath = file.join local cleanpath = resolvers.cleanpath Then you can something: \starttext \component toppath:/subpath/somefile.tex \stoptext All given that the top of the inputstack is still ok. If needed I can push/pop some relative code to start/stopcomponent alike code, in which case we will use a different prefix (no clue what name to use) Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
It would be REALLY NICE if the functionality of "pathrelativetome" was provided by ConTeXt.
in file-job.lua add around the function 'process':
You make me really happy, Hans! :-) I will change file-job.lua on my system to test. I can do this only on Wednesday.
local function toppath() [...] local pathname = dirname(inputstack[#inputstack] or "")
I didn't check it in lua. But the unix command "dirname" returns "." if given an empty string. That is, $ dirname '' . I don't know when "inputstack" could be empty... but maybe a warning should be issued, and the return value of toppath should be declared undefined. That is, it would be an error to call toppath if "inputstack" is empty.
[...] resolvers.toppath = topath
toppath? (double "p")
Then you can something: [...] \component toppath:/subpath/somefile.tex
Nice.
All given that the top of the inputstack is still ok.
Maybe the code could have some sort of "assertion" to 1. check for "stack underflow"; and/or 2. check if the file "somefile.tex" exists in the directory being popped from the stack. Just for "bug catching"...
If needed I can push/pop some relative code to start/stopcomponent alike code, in which case we will use a different prefix (no clue what name to use)
I don't think it is needed, since the toppath mechanism is rather generic. Writing "\component toppath:/subpath/somefile.tex" is fine by me. Unless you think that "start/stopcomponent" should be deprecated (or, at least, discouraged) in favour of this new version. Maybe it would be nice to have a function that returns the value of "inputstack[#inputstack]". But I don't know if it is useful... Thank you very very much for your time! André Caldas.
On 15-10-2012 21:42, Andre Caldas wrote:
I don't know when "inputstack" could be empty... but maybe a warning should be issued, and the return value of toppath should be declared undefined. That is, it would be an error to call toppath if "inputstack" is empty.
you should test \component x.tex \component y.tex and see if when y is loaded we can still get the right path
1. check for "stack underflow"; and/or
such things are dealt with
2. check if the file "somefile.tex" exists in the directory being popped from the stack. Just for "bug catching"...
it all happens before a file is loaded so the normal reporting is applied Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
2. check if the file "somefile.tex" exists in the directory being popped from the stack. Just for "bug catching"...
it all happens before a file is loaded so the normal reporting is applied
Actually, it is a bit annoying that \environment and \component do not report an error (or at least a warning message) when the file is not found. \usemodule at least issues a warning, but does not change the exit status code, so it is easy to miss the warning message. Aditya
2012-10-15 Aditya Mahajan:
2. check if the file "somefile.tex" exists in the directory being popped from the stack. Just for "bug catching"...
it all happens before a file is loaded so the normal reporting is applied
Actually, it is a bit annoying that \environment and \component do not report an error (or at least a warning message) when the file is not found.
+1 Marco
On 15-10-2012 23:27, Aditya Mahajan wrote:
2. check if the file "somefile.tex" exists in the directory being popped from the stack. Just for "bug catching"...
it all happens before a file is loaded so the normal reporting is applied
Actually, it is a bit annoying that \environment and \component do not report an error (or at least a warning message) when the file is not found. \usemodule at least issues a warning, but does not change the exit status code, so it is easy to miss the warning message.
I've added a message system > jobfiles > unknown tex file "what-a-mess" no detail about the kind of file (could be done but needs way more helpers as all now share common one) Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
1. check for "stack underflow"; and/or
such things are dealt with
By assert, I mean: assert(inputstack[#inputstack], "Input stack should never be empty when toppath() is called.")
2. check if the file "somefile.tex" exists in the directory being popped from the stack. Just for "bug catching"...
it all happens before a file is loaded so the normal reporting is applied
I meant an "assert()" inside toppath() and cleanpath(). But, as you said, this kind of "assertion" belongs to a "test case". André Caldas.
On 16-10-2012 00:26, Andre Caldas wrote:
1. check for "stack underflow"; and/or
such things are dealt with
By assert, I mean: assert(inputstack[#inputstack], "Input stack should never be empty when toppath() is called.")
normally if inputstack[#inputstack] == nil then assume path "." end is good enough ... (assert might make sense in situation where otherwise a crash would occur, but that never happens here) Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
1. check for "stack underflow"; and/or
such things are dealt with
By assert, I mean: assert(inputstack[#inputstack], "Input stack should never be empty when toppath() is called.")
normally
if inputstack[#inputstack] == nil then assume path "." end
is good enough ...
Is "inputstack[#inputstack] == nil" ever supposed to happen? When reading from stdin, maybe... then, it is ok.
(assert might make sense in situation where otherwise a crash would occur, but that never happens here)
If it is not supposed to happen, then it would indicate a bug, and should be detected as early as possible. A "crash" is a special case of that. But this is a matter of taste... :-) Cheers, André Caldas.
Hello!
It would be REALLY NICE if the functionality of "pathrelativetome" was provided by ConTeXt.
in file-job.lua add around the function 'process':
local function toppath() [...]
\component toppath:/subpath/somefile.tex
Works like a charm, Hans! Sorry I took so long to test... I copied file-job.lua from the GIT repository, and changed my tex files to use toppath:/path/to/file. For example: https://bitbucket.org/andrecaldas/math-video-classes/src/b743f22e4f567662ac5... Really nice! Thanks!! :-) Cheers, André Caldas.
Andreas wrote:
It would be REALLY NICE if the functionality of "pathrelativetome" was provided by ConTeXt.
Hans wrote:
in file-job.lua add around the function 'process':
local function toppath() [...]
\component toppath:/subpath/somefile.tex
Andreas wrote:
Works like a charm, Hans! Sorry I took so long to test... I copied file-job.lua from the GIT repository, and changed my tex files to use toppath:/path/to/file.
Will this be in the core, or is it just a fix for Andreas? I.e: should we document toppath() as a core command, or as a snippet? (Not that we have a general space on the wiki for snippets yet, but we need one and we will have one. Somethine in the next two weeks I'll write a "State of the Wiki" laying out what we have, what we don't have, and what we want next.) --Sietse
On 11/3/2012 10:13 AM, Sietse Brouwer wrote:
Andreas wrote:
It would be REALLY NICE if the functionality of "pathrelativetome" was provided by ConTeXt.
Hans wrote:
in file-job.lua add around the function 'process':
local function toppath() [...]
\component toppath:/subpath/somefile.tex
Andreas wrote:
Works like a charm, Hans! Sorry I took so long to test... I copied file-job.lua from the GIT repository, and changed my tex files to use toppath:/path/to/file.
Will this be in the core, or is it just a fix for Andreas? I.e: should we document toppath() as a core command, or as a snippet? (Not that we have a general space on the wiki for snippets yet, but we need one and we will have one. Somethine in the next two weeks I'll write a "State of the Wiki" laying out what we have, what we don't have, and what we want next.)
in the core .. it's a prefix just look in data-pre.lua for some more of them ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
Hello!
local function toppath() [...] resolvers.toppath = topath
There is a bug here. It should be resolvers.toppath = toppath (double "p" in the right side) André Caldas. -- - Por que altera a ordem natural da conversação! - Por que não? - Eu não gosto, não. - Você gosta quando copiam a mensagem original ao final do e-mail?
On 11/7/2012 8:34 PM, Andre Caldas wrote:
Hello!
local function toppath() [...] resolvers.toppath = topath
There is a bug here. It should be resolvers.toppath = toppath (double "p" in the right side)
In your local copy I presume, not here. ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
Hello, Hans! I quoted this from your message:
local function toppath() [...] resolvers.toppath = topath
When you commited version "eb24ffd", the following bug was introduced:
There is a bug here. It should be resolvers.toppath = toppath (double "p" in the right side)
In your local copy I presume, not here.
Because of the bug (I presume), you removed the whole code. git diff -r eb24ffd tex/context/base/file-job.lua [...] @@ -553,22 +553,6 @@ function resolvers.jobs.currentenvironment() return topofstack(v_environment) en local done = { } local tolerant = false -- too messy, mkii user with the wrong sructure should adapt -local function toppath() - local pathname = dirname(inputstack[#inputstack] or "") - if pathname == "" then - return "." - else - return pathname - end -end - -resolvers.toppath = topath ===> HERE IS THE BUG (right above) <=== - -resolvers.prefixes.toppath = function(str) - local fullname = cleanpath(joinpath(toppath(),str)) - return fullname -end Are we giving up the feature? I was kind of using it already... :-( Cheers, André Caldas.
On 11/8/2012 1:14 PM, Andre Caldas wrote:
Hello, Hans!
I quoted this from your message:
local function toppath() [...] resolvers.toppath = topath
When you commited version "eb24ffd", the following bug was introduced:
There is a bug here. It should be resolvers.toppath = toppath (double "p" in the right side)
In your local copy I presume, not here.
Because of the bug (I presume), you removed the whole code. git diff -r eb24ffd tex/context/base/file-job.lua [...] @@ -553,22 +553,6 @@ function resolvers.jobs.currentenvironment() return topofstack(v_environment) en local done = { } local tolerant = false -- too messy, mkii user with the wrong sructure should adapt
-local function toppath() - local pathname = dirname(inputstack[#inputstack] or "") - if pathname == "" then - return "." - else - return pathname - end -end - -resolvers.toppath = topath
===> HERE IS THE BUG (right above) <===
- -resolvers.prefixes.toppath = function(str) - local fullname = cleanpath(joinpath(toppath(),str)) - return fullname -end
Are we giving up the feature? I was kind of using it already... :-(
you have a messed up system ... the code is in data-pre.lua Gabs ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------
Are we giving up the feature? I was kind of using it already... :-(
you have a messed up system ... the code is in data-pre.lua
Indeed! :-) Sorry for the mess! Now, I have a bug at data-pre.lua. The variable inputstack is not defined there. Here is a fix: diff --git a/tex/context/base/data-pre.lua b/tex/context/base/data-pre.lua index 40b430b..3ef8c59 100644 --- a/tex/context/base/data-pre.lua +++ b/tex/context/base/data-pre.lua @@ -91,6 +91,7 @@ prefixes.home = function(str) end local function toppath() + local inputstack = resolvers.inputstack local pathname = dirname(inputstack[#inputstack] or "") if pathname == "" then return "."
participants (5)
-
Aditya Mahajan -
Andre Caldas -
Hans Hagen -
Marco Patzer -
Sietse Brouwer