Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/fontforge/libuninameslist

A library with a large (sparse) array mapping each unicode code point to the annotation data for it provided in http://www.unicode.org/Public/UNIDATA/NamesList.txt
https://github.com/fontforge/libuninameslist

Last synced: about 7 hours ago
JSON representation

A library with a large (sparse) array mapping each unicode code point to the annotation data for it provided in http://www.unicode.org/Public/UNIDATA/NamesList.txt

Awesome Lists containing this project

README 

        
LibUniNamesList

***************



LibUniNamesList holds www.unicode.org Nameslist.txt data which can be useful

for programs that need Unicode "Names", "Annotations" and block definitions.



Accessing LibUniNamesList Information

*************************************



Older versions of LibUniNamesList consisted of static data arrays which can

be accessed simply as arrays. This access still exists for compatibility

with older programs that use this method. One example program using this

older method is FontForge 20120731-b. The latest version of LibUniNamesList

contains functions which help simplify access to the same data.



Functions to access data in LibUniNamesList is listed here with the oldest

listed first, and latest added function listed last. If more functions are

added, they will be added to the bottom of this list.



1) const char *uniNamesList_name(unsigned long uni);

2) const char *uniNamesList_annot(unsigned long uni);

3) const char *uniNamesList_NamesListVersion(void);

These functions are available in libuninameslist-0.4.20140731 and higher

4) int uniNamesList_blockCount(void);

5) int uniNamesList_blockNumber(unsigned long uni);

6) long uniNamesList_blockStart(int uniBlock);

7) long uniNamesList_blockEnd(int uniBlock);

8) const char *uniNamesList_blockName(int uniBlock);

These functions are available in libuninameslist-20180408 and higher

9) int uniNamesList_names2cnt(void);

10) long uniNamesList_names2val(int count);

11) int uniNamesList_names2getU(unsigned long uni);

12) int uniNamesList_names2lnC(int count);

13) int uniNamesList_names2lnU(unsigned long uni);

14) const char *uniNamesList_names2anC(int count);

15) const char *uniNamesList_names2anU(unsigned long uni);

These functions are available in libuninameslist-20200413 and higher

16) const char *uniNamesList_Languages(unsigned int lang);

17) const char *uniNamesList_NamesListVersionAlt(unsigned int lang);

18) const char *uniNamesList_nameAlt(unsigned long uni, unsigned int lang);

19) const char *uniNamesList_annotAlt(unsigned long uni, unsigned int lang);

20) int uniNamesList_nameBoth(unsigned long uni, unsigned int lang, const char **str0, const char **strl);

21) int uniNamesList_annotBoth(unsigned long uni, unsigned int lang, const char **str0, const char **str1);

22) int uniNamesList_blockCountAlt(unsigned int lang);

23) long uniNamesList_blockStartAlt(int uniBlock, unsigned int lang);

24) long uniNamesList_blockEndAlt(int uniBlock, unsigned int lang);

25) const char *uniNamesList_blockNameAlt(int uniBlock, unsigned int lang);

26) int uniNamesList_blockNumberBoth(unsigned long uni, unsigned int lang, int *bn0, int *bn1);



For a better description of each function, check "uninameslist.h", and see

an example of how to use these functions in FontForge ~ 20140101 or later.



Alternatively, look at tests/call-test.c on how to call these functions.



Building and Installing LibUniNamesList

***************************************



Read INSTALL for details on building LibUniNamesList using this package.

This package contains 2 libraries, but only 1 is built by default:

	1) libuninameslist - holds NamesList.txt data from www.unicode.org

	2) libuninameslist-fr - currently holds French 13.0 NamesList.txt

If you need the libuninameslist-fr library then you will also need to enable

building it since the current default is not to build it. This will also be

recognized in functions 16..26 listed above as lang=1 (English=default=0).

	$ ./configure

	$ make

	$ su

	# make install (or use 'make install-strip' for smaller libraries).

	# exit

	$



Alternatively, if you do not have autoconf or automake available with your

computer, you can fetch the configure-ready libuninameslist-dist version.



Building and Installing LibUniNamesList and LibUniNamesList-fr

**************************************************************



The French library was built with the intention of maintaining backwards

compatibility with older programs (like FontForge 20120731), so it is a

separate library, but in this case, it is also linked to the base library.

If you ran 'configure' and 'make' earlier, you need to run 'make clean'

to remove the earlier libuninameslist (which won't link to this library).



NOTE: You may need to run 'make clean' if you ran "./configure; make" earlier.

	$ ./configure --enable-frenchlib

	$ make clean

	$ make

	$ su

	# make install (or use 'make install-strip' for smaller libraries).

	# exit

	$



Some platforms will work fine with the commands above, while others will

have trouble with not finding the French library because it does not exist

yet in your libraries. Here are three different variations:

Linux install (recent x86/amd64 distro):

	$ ./configure --enable-frenchlib

	$ make check

	$ sudo make install

	$



Linux install (recent s390x distro):

	$ ./configure --enable-frenchlib

	$ make

	$ sudo make install

	$ make check

	$

This is a bit of a cheat - we install the French libuninameslist-fr so it

is found by our 'to_check' libuninameslist when it uses -luninameslist-fr



Linux install (recent arm64 distro):

	$ ./configure --enable-frenchlib

	$ make

	$ sudo make install

	$ make check

	$

This fails make because of not finding the French libuninameslist, but will

pass 'make check' after both libraries are installed.



libuninameslist needs further improvements to avoid the above 'make' quirk,

but both libraries appear usable after being built and installed.



Added 'uninameslist.py' Python Wrapper

**************************************



A python wrapper is provided for users interested in libuninameslist access

using python. The easiest and simplest method for users to add it is to run

the install in two steps, first build and install libuninameslist, then use

python to install the wrapper:

	$ ./configure (may need --/prefix=/usr - use --help to see options)

	$ make clean

	$ make

	$ su

	# make install

	# cd py

	# python setup.py install

	# exit

	$



The build system can optionally also build installable wheels of the package.

To do this, pass --enable-pylib. Optionally, also set the `PYTHON` environment

variable to configure which python to use. The configured python must have pip,

setuptools and the wheel packages installed.



	$ PYTHON=python2 ./configure --enable-pylib (may need --/prefix=/usr)

	$ make

	$ su

	# make install

	# pip install py/dist/*.whl

	# exit

	$



Normally, a user default library install is made into '/usr/local', but for

some operating systems, you may need you to add '--prefix=/usr' if you will

be adding python access to uninameslist.py



This two-step install is necessary to avoid using AM_PATH_PYTHON() as part

of './configure' which interferes with installs that cannot include python.



Distro maintainers may also want to look at Adelie which has a great install:

https://code.foxkit.us/adelie/packages/blob/master/user/libuninameslist/APKBUILD



Default Build Process

*********************



The default build process can be summarized quickly like this:



	Download LibUniNamesList*.tar.gz

	Move it somewhere appropriate (temp perhaps?) and type:

	$ tar -xzf libuninameslist*.tar.gz

	$ cd libuninameslist????????

	$ autoreconf -i

	$ automake

	$ ./configure

	$ make

	$ su

	# make install-strip

	# exit

	$



The above steps work for both release or latest versions of libuninameslist.

Release versions provided by your distro may include a configure script which

allows you to skip extra package requirements (for autotools, automake), and

you would skip steps autoreconf -i, automake above, and start at ./configure



To see what options are available before running make, type



	$ ./configure --help



NOTE: Some Distros and Operating Systems may require you to run 'ldconfig' to

recognize LibUniNamesList if you are not rebooting your computer first before

loading another program that depends on LibUniNamesList. To do this, you may

need to run 'ldconfig' in 'su -' mode after you have done 'make install':



	$ su -

	# ldconfig

	# exit

	$



How to Update LibUniNamesList (for maintainers)

***********************************************



The main purpose for LibUniNamesList is to contain a current copy of the

latest updated version of Nameslist.txt as published by www.unicode.org

which can then be accessed by users needing this information.



This means every now and then, LibUniNamesList needs to be updated with new

information. To do this, you will need to get the latest NamesList.txt file

from www.unicode.org which will then be used to build a new namelist.c file.



Let's say the next version of NamesList is version 8.0.

1) Update configure.ac with the new timestamp:

	uninameslist_package_stamp [20150701]	<-"YYYYMMDD" format

2) Update configure.ac with the new NamesList version:

	uninameslist_nameslist_ver [8.0]

3) Increment the point release, was 0.4, now becomes 0.5. NOTE: This minor

   point number may get large since LibUniNamesList is backwards compatible.

	uninameslist_major_version [0]

	uninameslist_minor_version [5]

   (if a future update of ListeDesNoms.txt >7.0 appears, then update this;

	fr_major_version [0]

	fr_minor_version [4] )

4) Update the config system, and create a versioned buildnameslist.h:

	autoreconf -i -Wall

	automake --foreign

	./configure

5) Build the buildnameslist program:

	make

6) Import "NamesList.txt" and "ListeDesNoms.txt" into the same directory:

	make NamesList.txt

	make ListeDesNoms.txt

7) Create new updated {nameslist*.c, uninameslist*.h} files:

	./buildnameslist

	make clean

	make



Some operating systems have older or incomplete autoconf, automake, autotools

and it is best to provide a release that already contains ./configure without

needing autotools.



To build a release version containing ./configure it is best to clone a copy

into a fresh directory, and then remove all hidden files and directories to

reduce the final release size. You should also fetch the latest config.guess

and config.sub because your computer may have an older version - example:



	$ git tag 20210626

	$ git push origin

	$ git push origin --tags

	$ cd ~

	$ cp -pR libuninameslist_master libuninames_release

	$ cd libuninames_release

	$ rm -fR .git .github .appveyor.yml .travis.yml

	$ wget -O config.guess 'https://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.guess;hb=HEAD'

	$ wget -O config.sub 'https://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.sub;hb=HEAD'

	$ autoreconf -i

	$ automake --foreign

	$ ./configure

	$ make dist

	$ make distcheck

	$ ls -l libuninameslist*.tar.gz



When done, there will be a *.tar.gz file created in your directory. To avoid

confusing the downloadable github version from the release version containing

./configure it is best to rename this *.tar.gz file with the added '-dist'



	$ mv libuninameslist-20210626.tar.gz libuninameslist-dist-20210626.tar.gz



Added note: Going from Unicode 8.0 to 9.0, both of these libraries listed are

now bumped to 1.0 since this can be seen as a large change due to the library

compilation of the CRA being confused with Major Minor versioning.

Backwards compatibility for accessing data still remains intact for programs

that require access using older methods.

Users running older programs can recompile them to use the newer libraries or

optionally use a link pointing to the new libraries if cannot recompile them.



Adding And/Or Updating Languages (for translators)

**************************************************



If you are interested in adding a new language but don't know where to start,

then first recommendation is get a hold of the latest NamesList.txt file from

www.unicode.org and start with that first. You will want to use some sort of

plain text editor able to edit UTF-8 style files.

If you are interested in updating an existing file, there are some references

pointing to where latest NamesList style files were found, see in Makefile.am

or other nameslist*.c or uninameslist*.h for possible information if not seen

on readme files. Ask originating Authors if they want/need help if looking at

updating existing file (this is a large file, but can be done progressively).

If you are aware of translations in another format, they could be considered,

and functions 16 to 26 are available for displaying alternate languages.



Old Bugs Out, New Bugs In

*************************



If you have user type questions, a good place to look, or ask is:

	https://sourceforge.net/p/fontforge/mailman/fontforge-users/

	[email protected]



Please report any bugs, patches and/or improvements to:

	[email protected]



This file was last updated:

	2021-Jun-26, by Joe Da Silva