# Different ways to collect an extension's functions. - Thread: https://externals.io/message/48768 - Original archive: https://news-web.php.net/php.internals/48768 - Started by Richard Quadling on 2010-06-14T10:38:51+00:00 - Messages: 6 - Reply by email to internals@lists.php.net - More for agents: search with https://externals.io/search.md?q={query} · JSON API at https://externals.io/api · documentation at https://externals.io/llms.txt --- ## #48768 — Richard Quadling — 2010-06-14T10:38:51+00:00 URL: https://externals.io/message/48768#48768 Hi. I'm looking into bug#52079 (missing lcfirst function from function index). The cause of the bug is that the genfunclist / genfuncindex scripts haven't been run in a while. Whilst running them and fixing a small issue with ereg => preg, I notice that in the results of genfunclist, there are 3 different ways to collect an extensions functions ... 1) The majority use const (79 in total). # const zend_function_entry apache_functions[] = { # const zend_function_entry apache_functions[] = { # const zend_function_entry basic_functions[] = { # const zend_function_entry bcmath_functions[] = { # const zend_function_entry birdstep_functions[] = { ... # const zend_function_entry xsl_functions[] = { # const zend_function_entry zend_funcs_aggregate[] = { # const zend_function_entry zend_funcs_arrayaccess[] = { # const zend_function_entry zend_funcs_iterator[] = { # const zend_function_entry zend_funcs_serializable[] = { 2) The following use static. # static zend_function_entry disabled_function[] = { # static zend_function_entry mysqlnd_functions[] = { # static zend_function_entry php_sqlite3_class_methods[] = { # static zend_function_entry php_sqlite3_result_class_methods[] = { # static zend_function_entry php_sqlite3_stmt_class_methods[] = { # static zend_function_entry spl_funcs_SplFixedArray[] = { 3) The following don't use const or static. # zend_function_entry enchant_functions[] = { # zend_function_entry fileinfo_functions[] = { # zend_function_entry intl_functions[] = { # zend_function_entry litespeed_functions[] = { # zend_function_entry phar_exception_methods[] = { # zend_function_entry phar_functions[] = { # zend_function_entry php_archive_methods[] = { # zend_function_entry php_entry_methods[] = { # zend_function_entry php_oci_functions[] = { Regards, Richard. -- ----- Richard Quadling "Standing on the shoulders of some very clever giants!" EE : http://www.experts-exchange.com/M_248814.html EE4Free : http://www.experts-exchange.com/becomeAnExpert.jsp Zend Certified Engineer : http://zend.com/zce.php?c=ZEND002498&r=213474731 ZOPA : http://uk.zopa.com/member/RQuadling --- ## #48769 — Melanie Rhianna Lewis — 2010-06-14T10:56:57+00:00 URL: https://externals.io/message/48768#48769 · In reply to: #48768 My feeling is that they should all be static consts unless explicitly accessed outside of the object module or changed. Since typically they're a list of function/method entries only accessed by a class or zend_module_entry that is really the case. Without the static we're polluting the name space. Melanie > Hi. > > I'm looking into bug#52079 (missing lcfirst function from function index). > > The cause of the bug is that the genfunclist / genfuncindex scripts > haven't been run in a while. > > Whilst running them and fixing a small issue with ereg => preg, I > notice that in the results of genfunclist, there are 3 different ways > to collect an extensions functions ... > > 1) The majority use const (79 in total). > > # const zend_function_entry apache_functions[] = { > # const zend_function_entry apache_functions[] = { > # const zend_function_entry basic_functions[] = { > # const zend_function_entry bcmath_functions[] = { > # const zend_function_entry birdstep_functions[] = { > ... > # const zend_function_entry xsl_functions[] = { > # const zend_function_entry zend_funcs_aggregate[] = { > # const zend_function_entry zend_funcs_arrayaccess[] = { > # const zend_function_entry zend_funcs_iterator[] = { > # const zend_function_entry zend_funcs_serializable[] = { > > > 2) The following use static. > > # static zend_function_entry disabled_function[] = { > # static zend_function_entry mysqlnd_functions[] = { > # static zend_function_entry php_sqlite3_class_methods[] = { > # static zend_function_entry php_sqlite3_result_class_methods[] = { > # static zend_function_entry php_sqlite3_stmt_class_methods[] = { > # static zend_function_entry spl_funcs_SplFixedArray[] = { > > > 3) The following don't use const or static. > > # zend_function_entry enchant_functions[] = { > # zend_function_entry fileinfo_functions[] = { > # zend_function_entry intl_functions[] = { > # zend_function_entry litespeed_functions[] = { > # zend_function_entry phar_exception_methods[] = { > # zend_function_entry phar_functions[] = { > # zend_function_entry php_archive_methods[] = { > # zend_function_entry php_entry_methods[] = { > # zend_function_entry php_oci_functions[] = { > > > Regards, > > Richard. > > -- > ----- > Richard Quadling > "Standing on the shoulders of some very clever giants!" > EE : http://www.experts-exchange.com/M_248814.html > EE4Free : http://www.experts-exchange.com/becomeAnExpert.jsp > Zend Certified Engineer : http://zend.com/zce.php?c=ZEND002498&r=213474731 > ZOPA : http://uk.zopa.com/member/RQuadling --- ## #48770 — Gustavo Lopes — 2010-06-14T11:28:30+00:00 URL: https://externals.io/message/48768#48770 · In reply to: #48769 On Mon, 14 Jun 2010 11:56:57 +0100, Melanie Rhianna Lewis wrote: > My feeling is that they should all be static consts unless explicitly > accessed outside of the object module or changed. Since typically > they're a list of function/method entries only accessed by a class or > zend_module_entry that is really the case. Without the static we're > polluting the name space. > The static keyword is unnecessary, const already makes the variable have internal linkage by default. -- Gustavo Lopes --- ## #48774 — Melanie Rhianna Lewis — 2010-06-14T13:32:25+00:00 URL: https://externals.io/message/48768#48774 · In reply to: #48770 I know the ANSI C standard pretty well but I don't remember that at all so rather than just accepting it I checked it on my Mac using a small example. Compiler details are below: kring:Development melanie$ gcc --version i686-apple-darwin10-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5657) In my example static const did have internal linkage, const on its own in the global scope did not. You still need static. Melanie > On Mon, 14 Jun 2010 11:56:57 +0100, Melanie Rhianna Lewis > wrote: > > > […] > > The static keyword is unnecessary, const already makes the variable have internal linkage by default. > > -- > Gustavo Lopes --- ## #48775 — Patrick ALLAERT — 2010-06-14T14:01:38+00:00 URL: https://externals.io/message/48768#48775 · In reply to: #48774 2010/6/14 Melanie Rhianna Lewis : > I know the ANSI C standard pretty well but I don't remember that at all so rather than just accepting it I checked it on my Mac using a small example.  Compiler details are below: > > kring:Development melanie$ gcc --version > i686-apple-darwin10-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5657) > > In my example static const did have internal linkage, const on its own in the global scope did not.  You still need static. Right, "const" is not similar to "static const"! +1 for static consts for all of them Patrick > Melanie > > > > > > […] --- ## #48776 — Gustavo Lopes — 2010-06-14T14:12:28+00:00 URL: https://externals.io/message/48768#48776 · In reply to: #48774 On Mon, 14 Jun 2010 14:32:25 +0100, Melanie Rhianna Lewis wrote: > I know the ANSI C standard pretty well but I don't remember that at all > so rather than just accepting it I checked it on my Mac using a small > example. Compiler details are below: > > kring:Development melanie$ gcc --version > i686-apple-darwin10-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5657) > > In my example static const did have internal linkage, const on its own > in the global scope did not. You still need static. > You're right. I'd checked it here: http://publib.boulder.ibm.com/infocenter/iadthelp/v7r0/index.jsp?topic=/com.ibm.etools.iseries.langref.doc/as400clr16.htm and some other place, but at least the C99 standard says otherwise. -- Gustavo Lopes