diff options
Diffstat (limited to 'src')
| -rw-r--r-- | src/modules/rkt/rkt_scener.c | 121 |
1 files changed, 57 insertions, 64 deletions
diff --git a/src/modules/rkt/rkt_scener.c b/src/modules/rkt/rkt_scener.c index 11be927..fd08f99 100644 --- a/src/modules/rkt/rkt_scener.c +++ b/src/modules/rkt/rkt_scener.c @@ -70,6 +70,7 @@ typedef struct rkt_scener_t { til_setting_t *cur_edited; unsigned replacement:1; /* set when editing / replacing scener->scene */ unsigned editing:1; /* set when editing */ + unsigned refinalizing:1; /* set when fixing up an EINVAL @ finalize */ } new_scene; } rkt_scener_t; @@ -440,6 +441,7 @@ static int rkt_scener_handle_input_editscene(rkt_context_t *ctxt) scener->new_scene.cur_edited = NULL; scener->new_scene.replacement = 1; scener->new_scene.editing = 1; + scener->new_scene.refinalizing = 0; scener->state = RKT_SCENER_FSM_SEND_NEWSCENE_SETUP; return 0; @@ -484,6 +486,7 @@ static int rkt_scener_handle_input_newscene(rkt_context_t *ctxt) scener->new_scene.cur_edited = NULL; scener->new_scene.replacement = 0; scener->new_scene.editing = 0; + scener->new_scene.refinalizing = 0; scener->state = RKT_SCENER_FSM_SEND_NEWSCENE_SETUP; return 0; @@ -1047,24 +1050,8 @@ int rkt_scener_update(rkt_context_t *ctxt) r = rkt_scene_module_setup(scener->new_scene.settings, &scener->new_scene.cur_setting, &scener->new_scene.cur_desc, - NULL); /* res_setup deliberately left NULL for two reasons: - * 1. prevents finalizing (path is "...WIP-new-scene...") - * 2. disambiguates -EINVAL errors from those potentially - * returned while finalizing/baking into a til_setup_t. - * - * This way if such an error is returned, we can expect - * res_setting and res_desc to refer to /what's/ invalid, - * and correct it before retrying. - * - * It'd be rather obnoxious with how the setup funcs are - * implemented currently to try force returning the relevant - * setting+desc for errors found during baking... I still - * need to firm up and document setup_func error semantics, and - * ideally it would be nice if the finalizing could say what - * setting/desc is relevant when say.. doing an atoi(). - * That will require a bit of churn in the settings get_value - * API and changing all the setup_funcs to be setting-centric - * instead of the current char* value-centric mode TODO + NULL); /* res_setup deliberately left NULL to prevent finalizing + * while path is still "...WIP-new-scene..." */ if (r < 0) { /* something went wrong... */ @@ -1159,55 +1146,57 @@ int rkt_scener_update(rkt_context_t *ctxt) */ /* these *may* move into helpers, so scoping them as ad-hoc unnamed functions for now */ - if (!scener->new_scene.replacement) { /* expand scenes settings */ - til_settings_t *scenes_settings; - til_setting_t *scene_setting; - char *label, *as_arg; - int r; - - /* now that we know settings is completely valid, expand scenes and get everything wired up */ - as_arg = til_settings_as_arg(scener->new_scene.settings); - if (!as_arg) - return rkt_scener_err_close(scener, ENOMEM); + if (!scener->new_scene.refinalizing) { + if (!scener->new_scene.replacement) { /* expand scenes settings */ + til_settings_t *scenes_settings; + til_setting_t *scene_setting; + char *label, *as_arg; + int r; + + /* now that we know settings is completely valid, expand scenes and get everything wired up */ + as_arg = til_settings_as_arg(scener->new_scene.settings); + if (!as_arg) + return rkt_scener_err_close(scener, ENOMEM); - scenes_settings = ((rkt_setup_t *)ctxt->til_module_context.setup)->scenes_settings; - scene_setting = til_settings_add_value(scenes_settings, NULL, as_arg); - if (!scene_setting) - return rkt_scener_err_close(scener, ENOMEM); + scenes_settings = ((rkt_setup_t *)ctxt->til_module_context.setup)->scenes_settings; + scene_setting = til_settings_add_value(scenes_settings, NULL, as_arg); + if (!scene_setting) + return rkt_scener_err_close(scener, ENOMEM); - r = til_setting_desc_new(scenes_settings, - &(til_setting_spec_t){ - .as_nested_settings = 1, - }, - &scene_setting->desc); - if (r < 0) /* FIXME TODO we should probably drop the half-added value??? */ - return rkt_scener_err_close(scener, r); + r = til_setting_desc_new(scenes_settings, + &(til_setting_spec_t){ + .as_nested_settings = 1, + }, + &scene_setting->desc); + if (r < 0) /* FIXME TODO we should probably drop the half-added value??? */ + return rkt_scener_err_close(scener, r); - r = til_settings_label_setting(scenes_settings, scene_setting, &label); - if (r < 0) /* FIXME TODO we should probably drop the half-added value??? */ - return rkt_scener_err_close(scener, r); + r = til_settings_label_setting(scenes_settings, scene_setting, &label); + if (r < 0) /* FIXME TODO we should probably drop the half-added value??? */ + return rkt_scener_err_close(scener, r); - r = til_settings_set_label(scener->new_scene.settings, label); - free(label); - if (r < 0) /* FIXME TODO we should probably drop the half-added value??? */ - return rkt_scener_err_close(scener, r); + r = til_settings_set_label(scener->new_scene.settings, label); + free(label); + if (r < 0) /* FIXME TODO we should probably drop the half-added value??? */ + return rkt_scener_err_close(scener, r); - scene_setting->value_as_nested_settings = scener->new_scene.settings; + scene_setting->value_as_nested_settings = scener->new_scene.settings; - } else { /* simply replace current nested settings */ - til_settings_t *scenes_settings; - til_setting_t *scene_setting; + } else { /* simply replace current nested settings */ + til_settings_t *scenes_settings; + til_setting_t *scene_setting; - scenes_settings = ((rkt_setup_t *)ctxt->til_module_context.setup)->scenes_settings; - if (!til_settings_get_value_by_idx(scenes_settings, scener->scene, &scene_setting)) - return rkt_scener_err_close(scener, ENOENT); + scenes_settings = ((rkt_setup_t *)ctxt->til_module_context.setup)->scenes_settings; + if (!til_settings_get_value_by_idx(scenes_settings, scener->scene, &scene_setting)) + return rkt_scener_err_close(scener, ENOENT); - /* FIXME TODO: keep this around until finishing the context replacement, - * and put it back in all the failure cases before that point. - */ - til_settings_free(scene_setting->value_as_nested_settings); + /* FIXME TODO: keep this around until finishing the context replacement, + * and put it back in all the failure cases before that point. + */ + til_settings_free(scene_setting->value_as_nested_settings); - scene_setting->value_as_nested_settings = scener->new_scene.settings; + scene_setting->value_as_nested_settings = scener->new_scene.settings; + } } { /* finalize setup, create context, expand context scenes or replace existing */ @@ -1230,15 +1219,19 @@ int rkt_scener_update(rkt_context_t *ctxt) if (r != -EINVAL) return rkt_scener_err_close(scener, r); - /* FIXME TODO: error recovery needs a bunch of work, but let's not just - * hard disconnect when finalize finds invalid settings... especially - * with the addition of til_setting_t.nocheck / ':" prefixing, this is - * much more likely. + /* Invalid setting found @ finalize! go back to prompting for input */ + scener->new_scene.cur_invalid = scener->new_scene.cur_setting; + /* But since it's @ finalize, we're in this half-finished awkward state where scenes_settings + * has already been expanded with this settings instance hanging off it. So it's like the + * previous step just needs to be suppressed altogether - it's not a replacement scenario, + * and we definitely don't want to re-add again either. We just want to re-finalize after + * the cur_invalid has been fixed up. */ - scener->new_scene.settings = NULL; + scener->new_scene.refinalizing = 1; - return rkt_scener_send_error(scener, EINVAL, RKT_SCENER_FSM_SEND_SCENES); + return rkt_scener_send_error(scener, EINVAL, RKT_SCENER_FSM_SEND_NEWSCENE_SETUP_PROMPT); } + scener->new_scene.refinalizing = 0; /* have baked setup @ setup, create context using it */ r = til_module_create_context(setup->creator, |